Marble size sorting machine
HTML-код
- Опубликовано: 1 окт 2024
- There's not much of a market for high tolerance marbles, so marble size inconsistency is an actual problem for some projects. But aside from using a gravity-fed size sorter to organize your collection, this project is also a simple way to demonstrate some probability distributions! Yay!
No, seriously: the Bell curve shows up everywhere- the average size of a raindrop, how many kittens your cat is likely to have, the number of blades o' grass in a lawn, or the approximate amount of rainbows & elephants found in a country- anything that has more than one example can be thought of as a distribution. This type of statistical thinking is not easy for us, because it is often counterintuitive, but it is one of the most useful tools we have to help us make sense of a confusing world. Mathematics lets us see clearly by giving us what we need to make unbiased predictions; it is something that we can actually trust, even when our hearts or eyes may mislead us.
Construction notes:
The collection trays were not the best way to do this. I improvised them from scrap. I might recommend higher-backed boxes that have a bit more width. If there were 6 compartments instead of 8, I think it might be easier to use, because it wouldn't have such a narrow range per cell. Such a small width is sometimes difficult to calibrate. Just think of mine as a proof of concept, and improve on it. It's hard to justify making more than one of these, when this one works to solve the problem just fine.
Music:
"Silver Flame" Kevin MacLeod (incompetech.com)
Licensed under Creative Commons: By Attribution 3.0 License
creativecommons...
its really satisfying how the marbles appear to slow down before they drop, as if each marble has a final destination in mind and slows down in anticipation for its arrival in its new home
Thomas Kaldahl dude. This thing needs more likes.
I think it is neat that the marbles automatically slow down right before falling through.
No they don't! They're just following a curved path. Along and then down, which is longer. They are actually speeding up- watch the spin. Mindblow.
They still slow down with regards to horizontal movement, though, though they do not slow down entirely. ;P
I guess you're right; they slow down rapidly with respect to that. In fact, they stop making horizontal progress entirely once they hit the bin.
Haha. I was referring to the frame of reference of them on the sorting slide, as Earthbjorn stated, but that's true as well.
To see what's happening here, think of the part of the marble in contact with the tracks. It will roll on circles near the end of each side of the marble, right over the point of contact. As the bars move further apart, the circles shrink, since they get closer and closer to the ends of the marble, until finally they shrink away to nothing and the marble falls through. I diagram would be useful here, but I don't have one :-P
Anyway, the smaller the circles of rolling contact get, the faster the marble must spin to move at the same speed, just like a small wheel has to spin more than a big wheel to cover the same ground. Since friction is keeping the marble from slipping as it goes down the track, it has to spin faster and faster. At the end, right as the circle shrinks to nothing, the marble would have to spin infinitely fast, but of course it starts slipping eventually. Still, this is why the marble slows down before it falls - more and more of the momentum from rolling downhill is pulled in to angular momentum to make the marble spin faster and faster as the circle of contact shrinks away, until the marble very nearly stops right before it falls through.
The higher the friction between the marble and the track, the closer it will come to stopping. To demonstrate this, it would be cool to run some rubber balls down the track ;-)
You must have presorted the marbles in the beginning of the video. Right? It really had me thinking. There are other possibilities that would cause the same separation to occur, but I would assume them to be more unlikely.
Yep.
I was wondering the same thing.
zachell1991 same
pocket83 When you adjust the bars to further sort the selected "average sized marbles" you should take one of the smallest outlying marbles and set the top of the sorting device to the width of that marble. Now you'll be sorting the marbles more finely over the entire length of the device and should more closely recreate the theoretical curve. As it stands, the sorting only begins at about a quarter of the way down and so your results will be less accurate. This is just my suggestion. as I love the concept and your content is fantastic.
I like it when you get mathematical. Thanks for your hard work.
Thanks for the appreciation!
it's great to watch. he's also got great mechanical advice ;)
I just watched How it's made on Tinned Pineapple and guess which method they use to sort the pineapples to sizes :)
I instantly remembered your video!
Cheers!
Could very well be that the distribution isn't normal, depends entirely on the method that the marbles are made. If they rolled the molten glass between two steel rollers to form the sphere, there's a set upper boundary to the size, and the distribution would fall off from there in the smaller sizes, depending on the temperature of the glass or something else.
Not to say that that's how they do it or that there isn't a normal distribution, just that that's a possibility.
Good point. Similarly, when I adjusted the lower part of the device (and not the top), I only extended the variance of the new set in one direction, larger; there was, in this case, also a boundary... wait... technically, that's not right- since the parallel bars had a new, completely different slope after the adjustment. Who knows? Though provoking stuff, isn't it?
Think you are correct that you only extend the variance in one direction:
by adjusting the lower part of the device only, you apply a linear transformation to the set, with width at the very bottom changing by X, and width at the very top not changing at all. This would skew the entire distribution to the left, since all marbles will fall off later. The variance will increase, but strictly in one direction. (Since marbles cannon physically fall off sooner).
I'd like to see the same thing but with smaller bins to get finer resolution on sorting and make a better distribution "graph".
You should put a funel on the start where you can just put many marbles at once
The way they brake smoothly to a (near) stop before dropping is so satisfying :) Really cool device!
i saw some marbles push other marbles further down, is that a major problem?
Not often. The weird ones do jam the device sometimes. But you have to expect some anomalies, deformities, and unlikely events. That's the nature of it! You can always repeat the run to get better results.
Do one marble at a time. Only put the next one in once you see the previous one has fallen
Agreed. But patience is easier said than done!
You could add on a funnel system with a rotary paddle release mechanism at a certain tension so it only releases marbles as a steady drop rate, so no two marbles will interact on the guide, no?
pocket83 why seperate by size when they are all sizes are different colours?
I really like this. We human beings do have a hard time adjusting to statistical concepts and this does a great job of illustrating distributions.
I imagine if that you would have set the rails by choosing the largest of the large group and the smallest of the small group you might have ended up with an even more normal looking distribution?
Thanks Pocket!
As I am watching, I wonder why not make a hopper?
Simplicity.
This is awesome! What possible reason could someone have for disliking the video.
I think it's more of a case of them disliking _me._ I mean, no real human being dislikes watching pretty rolling marbles organize themselves, unless they've gone completely deranged.
Well as far as them being deranged, I concur. One might say, they have lost their marbles.
Perhaps it is beyond the scope of your expertise, but a video on carburetors would be very interesting. Such a delightful little mechanism in a small, inexpensive and mechanical package.
What if you made it so that the bars were movable (within a guide of course) and you set values to the compartments on the tray. Then give players marbles and have them try to get the highest score within 3 moves.
I imagine that the coordination required to move the ramp to keep the marble on the "ramp" would be challenging. The higher scores/values would be closer to the top.
Play would start the marble at the bottom and move the rails to get the marble closer to them. The physics of this would make an interesting video in itself.
How much variance is there in the process? If you put the same marble through 100 times, what's the spread on where they fall through?
Marbles each have their own personality. The more egg-shaped ones will give a wider spread. In general, consumer marbles with a reasonably high tolerance will fall into the same cell 7 or 8 times out of ten.*
This is a great question, because in projects where precision marbles are needed (yes, that's a thing for me), variance is a key consideration in the marbles that I select. My usual method for establishing a tolerance goes something like,
_this set must land in cell x at least five out of six times._
Variance has everything to do with roundness.
*Note that this is largely dependent on the distributor. In my experience, craft stores have the least consistent quality marbles, and, perhaps ironically, dollar stores usually have flawless gems mixed into their $1 per 100 ct 'variety' bags. Seriously--I've found like nearly optical quality marbles there in at least one case.
Ah it makes a lot of sense that it's about roundness. I guess it might end up depending a lot on the orientation it's placed down in. Like, it might end up rolling around its major axis and seem bigger, or around its minor axis and seem smaller, or somewhere in between. But there's no easy way I can think of to control for that
Your videos are so inspirational and thought-provoking, thanks for sharing!
Thanks for letting me know that you like it!
This could probably replace manual sorting labor with using electricity (if the system is refined).
it needs a hopper.
I could watch those marbles all day. I hope we're getting a pocket squared video on this later, wink wink
The opening 45 seconds was OH SO SATISFYING to an OCD'ish personality. :-)
Put a cone to do alot in time.
That’s neat! 😃
Semi automatic loading mechanism would be great for this, so you can let each marble sort without bumping into each other but increase sorting speed for big sets of marbles. Just some loading chamber, channel to line marbles and pushing leaver with rubber band to push it back to lock position. Hope you dont mind this input, love your idea, your loyal fan.
That would be cool. It would also be fun. I wanted to make it as simple as possible though.
I think (this is all theoretical) if you wanted to get your marbles into sizes that you actually want, and not have to adjust the rig many many times, you could use a known size to set the bars apart. Or, if you were moving from something with very precise tolerances (Ball bearings) to lower tolerance (Ebay/dollarstore marbles) you could set the bars for that known, working size.
So if you were designing something around, say, pachinko balls... But they are more expensive, and heavy since they are metal. So whatever the reason, you need to switch from them to marbles. Since the metal ball bearings are mass produced and require tight tolerances to work in the game, but marbles are not, you would know what size you want (11mm). So, by putting the top of the bars 10mm apart, and the bottom 12mm apart you could get your initial and highly refined sort very quickly. You would probably have to modify the design to kick out the super large marbles on the end, as you would probably have lots of them because of the initial sort being very precise.
Maybe I'm over-thinking it...
I got a bag of 25 kilograms of very cheap indian marbles for my Slingshot because I don't wanna poison my own yard with lead balls
So the big bag of marbles has a tolerance between 13 and 18 millimeter
That is a big tolerance 😂 I use two alloy profiles and put 6 bags under the rail
My kids love collecting and sorting my ammo 🍻
The first year engineering class at my university assigns a semester-long group project, where we have to design a marble sorter and barcode reader out of a Lego Mindstorms kit, which has some motors, and various types of sensors such as touch or light/color sensors, and program it with LabVIEW. The finished product should be able to read a barcode containing a prescription of certain size/color marbles to collect, and fill out the dosage from a pool of about 75 marbles. It was a very challenging project and this video made me think back to it and how crappy our design was
Wish I could use this concept to sort laundry.
In an earlier video, you stated you wouldn't be a good teacher in a classroom because you get to edit out your mistakes, but this video demonstrates that you could be a very helpful *online* teacher! Y'know, maybe something along the lines of Khan Academy, or a Pearson example video.
Watching the marbles get sorted is eye candy. Now what if you turned it into a display? Where the marbles get automatically re-fed into the top. I think I'd never get any work done...
Would be such a good teaching tool for statistics.
Now, how about a device that illustrated Pareto's Principle? Only like 25% joking, I think that would be awesome, though have no idea how it could be done.
I see possibilities... A multi run marble machine. Each run for a slightly different size marble. All starting on a single run and than being redistributed by size, than combined back together again just before the lift.
This is really cool, love your content! Could you try making some sort of web shooter using the two liquids 1 6 hexanediamine and sebacoyl chloride?
No. It smells too bad. And it burns.
So why dont marble making companies have this sorter? I guess it's because its not super important for most people who buy marbles but idk. Also the first run you can not find much of a slope but if you re run each section of the first run in order and add them in it would be more accurate and another better way would be to lower the decline on both the rails lower and collecting system plus lower the height of the drop then instead of a few dividers you could have alot of ridges to hold them in place with an average marbles distance between them so they are all in order basically from smallest to largest. The lessened height would just make it so the marbles wont bounce over the ridges or roll too far down the rails.
Anyway very interesting video, thanks for sharing this with everyone!
Edit: the second option would give a much better visualization of the average size but is passed the whole point of the size sorter as it would not be as simple to group them together because you would end up with many groups, which i guess would only be better for extremely precise projects
waw 😂
Upload more your vids are awesome
I'll try to have one next week. This stuff isn't easy to make, so bear with me!
About your question i think it's 3
i think it's 4 because the majoraty of the marbels is in tray 2 and there are alot more in tray 3 than tray 1 if that makes sence :)
Another great video!
It would be interesting to see this incorporated into a marble machine as a way to direct marbles onto different paths.
That would be spectacular. Imagine a double helix using two different colors. It would be mesmerizing to watch it self-sort and then spiral down around itself.
Now you should make a auto-feeder so you don't have to load by hand.
A reddit post I saw today showing a similar system for sorting limes: www.reddit.com/r/Damnthatsinteresting/comments/76hvq1/indian_farmer_sorts_lemons_by_their_size/
your logo is the octahedron of transcendence (Google it)
Yeah ... its cool looking. But come on now. You've set the bar way too high for yourself for this to be enough right? Now sort them by MASS or if you think you can pull it off... by color. If you can do color with no electrons I'll buy you lunch.
Easy to sort by color! Different color marbles will come from different batches, possibly from different lines or even different factories, and therefore have somewhat different sizes, just like you saw in the intro to the video. Of course, the problem comes when the difference between different color batches is smaller than the variance within a batch...
shouldn't be too hard having a funnel above the machine allowing you to dump a bunch of marbles and let it do its thing? your genius enough!
At 1:23 the 7th marble (2nd from the back) is the one to go inside the latest. You're welcome
Would you say it's a... mar-Bell curve?
Sorting marbles can be a start of a analog computer. Each marble representing an instruction. Ran trough the 'computer' one at the time.
Can you make a funnel that lets marbles pass through with a window gap of 1 second?
U should make a box on top of the sorting machine to put them in so u wouldnt have to put them in there yourself you could just dump them in the box.
This is clearly a color separater. You don't even know what you have made.
So that's what normal distribution is for, measuring marbles ;)
OMG where are we going with this ?I'll have to cut more paint cans open to get marbles lol
You should make a hopper so you can just dump a whole bag and sort them all real fast and with less effort.
But.....
why would I want to watch my crab?
this is real, its just the marbles and where it slides on is different sizes
Pocket should be a math teacher. He's also head and shoulders above bill Nye as a scientist
Pocket83 I've been thinking that it would be really cool if you built one of these on a on a larger scale: ruclips.net/video/skvnj67YGmw/видео.html. It would better illustrate the differences in efficiency.
The reason the marbles slow down before they drop is very easy to explain: Magic.
I could watch that marble sorting all day long, simple - yet fascinating!
Ha! I know what you mean. I found myself staring at it for a long time. That's why I left the opening footage so long.
I'm surprised there is so much variance in one type of supposed same marble.
For a moment I thought it was a video from jelle's marble runs
In the Gauss-Distribution modell all marbles have the same diameter and go all through the same opening.
Then they fall down through a bunch of nails (or similar). I have build this years ago for a demonstration in an institut in a University in Germany. The plan was in a mathematicsbook.
Not really, but i used the schematic and build a 1m (about 3 feet?) model. Was a nice project with many try and error.
Cheerio from Germany
That sounds like the original Pachinko machines. It can be used for gambling. In that case, you can have jackpots on the tail-ends of the curve, and the 'house' still wins.
Note: I don't gamble. Ever.
Olives are sorted this way. Probably means it's a very old method.
Am I the only one who didn’t know marbles were different diameters😂
Your "sink a nail with one hit" video was on PewDiePie today!!!!!
I came here looking for inspiration for a DIY machine to sort nuts by size. Fascinating watch. Thank you.
Thanks. That's a much harder problem. The first idea to spring to mind is a rotating, vibrating drum that has thickness-based slots which would allow smaller nuts to escape. Might need some oil. Same theme as this video: over time, math gets closer and closer; eventually, all the small nuts _will_ escape! I'd be interested to hear what you come up with.
That is clever, nice work sir!
Neat how they slow down before dropping, ....!
i belive it's 4
What is you’ve already lost all your marbles...? LOL
Being creative and logical is a curse, but watching you use both sides of your brain at the same time... WONDERFUL (I don't feel so alone now) I bet you drive other people crazy too, so much more to say but jumbles up... Thanks for the inspiration and a fresh breath of air
I tried this for a low-quality buckyballs neodymium magnet spheres. Couldn't find a suitable material for the rails though.
That's a tough one. Some type of fiberglass or carbon fiber rod? Honestly, imperfections in the material(s)* would be a much larger percentage of the sphere diameter, so this method would probably _not_ be the way to go. Tedious and frustrating caliper use might be more suitable, but they'll stick to that, as well.
I hate to say it, but you should buy better ones if they aren't crystalizing- some problems only have painful solutions :(
*random error
Can you make this much bigger to sort cans and tennisballs?
for some reason this reminds me of the "wave resonance" produced in the single particle double slit experiments.
I wonder what sort of tolerances can be achieved with this type of sorter. I've been playing with the idea of using this to sort airsoft bbs for use in 3d printed bearings.
I imagine the precision on the initial rods is what would be the main factor in the machines sorting capability coupled with the rigidity of those rods. Maybe some long precision ground shoulder bolts if there are any long enough... Oh! Perhaps linear bearing rods would work well!
Look for an old printer. The rods inside are flawlessly smooth and precise. With respect to tolerance, mine uses inexpensive zinc-plated rod, and it goes way beyond my visual acuity. If you really _need_ to know a numerical value, get back to me and I'll take some measurements. Good luck.
Thanks for the great videos - really interesting and educational. Thanks Rod
This is the quality content I love seeing from you.
Brilliant ideas as always :D
Hey, it's good to see a vid! It's funny...I have absolutely no intention of making a marble sorter. No need, no desire. But I thoroughly enjoyed the video! It's almost like sitting down with a friend and having a conversation about one of HIS interests. I don't have to have any "take away" from a conversation like that to enjoy it. So...thanks! I also appreciate the metric conversion ;)
To avoid the bending problem while adding a lot of precision of the sorting process, I imagine a multiple pair of bars builded like the original Donkey Kong game :D However, fine tuning this sorter could be tricky I think...
put a funnel on the thing and it's perfect
Very interesting. Have you seen how ping pong balls are spherically graded? They are rolled down a wide straight ramp towards 5 collection areas - the balls that go through the centre collection area (least deviance from a straight line of release) are graded 3*, the 2 areas either side of that are 1* and the rest are ungraded balls. I think this could be easily replicated for any projects where roundness in key.
Wow, that is really clever! But what if a malformed blob came out? It would just slide down the center of the ramp ;(
Note: grade -1* are probably used specially for carnival games!
I didn't see you address repeatability. I imagine most marbles aren't all that spherical, and so the same marble might fall into different bins on consecutive runs when your bars start approaching parallel. This might actually be a good technique to measure sphericity in your marbles too! (A more consistent binning at very tight tolerances suggests a more spherical marble)
Yes! Exactly.
DUDE, PEWDIEPIE just used your one hit nail video in his latest "THIS KID IS THE NEW PEWDIEPIE" @ 5:30
Pocket is famous
Can you build a machine that helps me find my marbles? Thanks.
Have you tried a hopper mechanism so you won’t have to do it one by one?
One thing about marbles is that they're generally not very perfect spheres. It would be interesting to take all the marbles that fall into one bin and run them through the sorter, unaltered, again, to see if they still land in the same bin.
Incidentally, when they sell ball bearings, they have grades according to how perfectly spherical they are. Found this out when I needed to get 3/8th inch balls for a marble machine.
Repeated trials *will* increase precision. There are certainly defects, but manufacturers are aiming at spherical in general, so they often are quite close. Still, there's a reasonable degree of quality that you will try to achieve with this gizmo, that is, a spot that rests within a certain variance. Once that sweet spot is found, the marbles that hit the same bin a few time in a row are extremely spherical; variations and anomalies will readily register as a different size occasionally.
All things considered, I'm more impressed by the sphericity that marble makers have achieved than their ability to sort within a tight size tolerance- not that most kids care what size they are while they are rolling the things around on the porch.
3.5, mathematically speaking.
You should make a hopper or funnel so you can put all the marbles in at once.
I knew it was by size since the start if you didn't know it's common sense
it's simple BUT AWESOME!
All it’s missing is a funnel/hopper at the top to be more efficient
Phân biệt theo kích cỡ bi
You should have a funnel at the top
Theo kích cỡ viên bi
Can someone explain why the marbles seem to stop before dropping in their bins ?
I read some of the comments, but I couldn't understand
This is not too tough, but you'll have to think about it for a bit ;) Imagine using your two index fingers to spin a basketball by its sides: it doesn't go forward at all!
Back to a marble rolling on rails. Think of two things: 1) where the marbles are contacting the tracks, and 2)how their paths are being curved from forward to downward. As the round tracks open up on the underside, the marbles spin more, yet move forward less. They are still spinning just as fast- but because they start to contact the rails more at their sides as they move down, the spinning doesn't get them as far. When they are just about to drop, they are spinning almost perfectly by their sides, like the basketball.
By analogy, a large tire travels farther with one revolution than a tiny tire does. Similarly, as one roll of a marble takes place over a smaller area, it will travel forward less.
I think you mean high-precision marbles, and that you have a low tolerance for imprecision.
I need more specific context to respond to that. However, there is a point of diminished return with precision in semantics; if an idea is successfully expressed, then the ends justify the means for mild misuses of the language. Too strict an adherence to form also breaches the absurd in places, especially with some nebulous cases, like the word 'precision,' where its precise definition is a bit of a stretch from the more generalized vernacular. Inflexibility in language is like having a rope that doesn't bend: it will quickly obsolesce in favor of something more dynamic.
tol·er·ance
_noun_
2. an allowable amount of variation of a specified quantity, especially in the dimensions of a machine or part.
Not that a person who has sufficient command of the language couldn't artfully use a noun as an adjective, anyhow. I haven't even had a morning cup of coffee yet. Sheesh. Cut me a break.
Just need a little bell for each marble and can play little melody
Very cool machine for sorting and statistics.
Why have you not been working with Wintergaten!?
Awesome approach to something that was nothing but laborious to learn in my AP Statistics class in high school. It's always a great pleasure when you do maths in your videos. :)