For the 2015 Hack a Day Prize, [Arkadi] is working on an educational platform using the Arduino for Smart Juggling Balls.
His goal for this is to create an open-source platform which, beyond juggling, can be used to teach physics in both an interactive and fun way. The juggling balls feature a RGB LED, an Arduino pro mini, and some MPU’s. They can be programmed to change color based on acceleration, gravity, centripetal force, rotation, or even controlled remotely. As the project develops further, he also plans on creating lecture content to go alongside the project, which would make it an excellent and interactive project for a high school (or even college) tech or physics class.
It’s not completely done yet but he’s already posted all the source code and instructions for making your own set over on GitHub. Stick around after the break to see the prototype balls in action and don’t forget to get cracking on your own Hackaday Prize entry!
Alternatively you could just juggle digital balls using the Microsoft Kinect and this clever hack…
20 thoughts on “Smart Juggling Balls”
Excellent! I was thinking about doing my led juggling balls for the hack a day prize last year but didn’t make the time for it:
It looks like Arkadi has got his project further along than mine. I never got ball shaped cases sorted out. These guys are worth mentioning https://www.youtube.com/watch?v=IMQMb8d_NK8 but I’m not sure if there’s open source.
keep up the good work . !!: :-)
its Certainly not the first version of glowing balls on the web. there was recently a campaign on idiegogo of something more progressed than my version: https://www.indiegogo.com/projects/juggglow-the-next-level-of-juggling
But i`m not aiming for a product. My interests are hobby level and educational. I really would like people to build their versions and by that improving the overall design in order to make something that will be possible to bring to a class, and perhaps motivating children to study Eng. and precise science by using a glowy and blinky thingy to show actual physics in real time.
Nice, I like it!.
WS2812B datasheet suggest that you should have 100n capacitor in parallel of Vdd and Vss for each WS2812B chip.
How does it distinguish between gravity and acceleration? Theoretically this is impossible. Look up “equivalence principle” on the Wikipedia.
If the object is also under acceleration, magnitude of the resulting vector would be different than 9.81m/s^2.
Since it doesn’t have it own propulsions, once it leaves source of acceleration e.g. your hand or a sling or a rubber band, then it is pretty much gravity from that point on.
True. It wouldn’t work on the Moon or Mars, then, but assuming the juggler is on the Earth is probably sufficiently general for most people ;-)
Hi, i have on board accelerator and a gyro so basically after you know the sensor orientation (see 6dof fusion) you can calculate the gravitational effect on the board. and that of course assuming you are on earth with a know gravitational force. did i miss something?
any way for my current application i just calculate the norm value (sqrt(ax^2+ay^2+az^2)) and use it as a reference for 0 gravity (in air), 1g and high g. that’s all that is needed for juggling. later i will improve it perhaps for something with rotation, gesture recognition and we will see..
>0 gravity (in air)
There is always gravity in flight. Even when the ball reaches the highest point with a zero vertical component velocity, gravity is still there acting on it pulling it down.
hah, sorry for misusing concepts. of course you are right. there is always gravity. i meant zero acceleration as measured by the on board accelerometer normalized to g scales (and here was inserted the misuse of the term gravity and reference coordinates :) ).
Getting WiFi telemetry data about its acceleration, velocity, spin etc and plotting them would be of some educational use in Physic than blinking LED.
Perhaps you missed this part out, but i have a simple nrf24l01 Transmitter on board, so transmitting the data out and showing graphs and plots is part of it. the blinky part is attention and motivation which is a very important factor in education (by my opinion)
Imagine if they all fed synchronized timestamped data to a PC with a webcam, and could overlay a rolling chart of the data from each sensor axis for each ball. Better yet, use OpenCV to track the relative 2d location (in arbitrary units or scale using photogrammetry) to show students the relationship between the position over time and the measured acceleration.
Yep, that’s part of the ideas. I bought Pixy CMUcam5 Sensor, which i thought is working like magic but the truth is that the camera is not that high a resolution. maybe raspberry pi + webcam + OpenCV will be better. in time i will test it out. or maybe someone else who will decide to work on the project too.
Regarding the synchronized time stamp from basic tests, its not that important. its enough to have a time stamp from each ball and the synchronization isn`t that important because the delay (from transmission) is less that the measurement frequency (i strive to use a 100hz reference which is enough for most things).
This is a really killer project. I can think of all kind of uses for this while learning some new patterns. You can actually analyze how solid you are. Now they just need to come up with some new smart balls that don’t roll under your bed when you drop.
These guys have been doing a very impressive similar version; they’ve been at it for a while: https://twitter.com/glowrbs
I’ve tried it a few months ago and was blown away :)
Thanks for the link. they look nice but i didn`t find any links to their files (they claim open source….)
Besides i have noticed they are using a lot of LEDs, i have another version which uses 52 LEDs at almost the same size but besides being really glowing and make an “Effect” its not practical. the batteries are drained out so fast that its no good (and i`m using 800 mah lithium batteries). i think the right amount is somewhere between 6-10 which is enough light and has pattern generation options. the juggglow product uses only 2 LEDs and as far as i have seen it barely glows at all.
the 52 LEDS video is here: https://www.youtube.com/watch?v=jQbHNXzZFqI
Hi – I’m a maker on the Glowrbs team. Cool project you’re working on!
Yeah, we have a lot of LEDs (~40) – for several reasons: first because we use them for pattern generation/animation, and second, to have a perfectly uniform color with absolutely no shaded areas. Battery wise we run with a 1Ah, but rarely power up the Glowrbs with all LEDs on max brightness: at 75% brightness it still gets us ~5h of lifetime. Regarding opensource we’re actually in the process of sharing the construction journey of the Glowrbs hardware on hackaday.io and github. So far, we’ve been pretty busy in the makerspace and have only been posting a few photos on instagram/twitter.
Impressive that you’ve made a 52 LED version! We’d love to see photos of it if you have a link to share :)
Hi, i don’t see an option to just add photos in a reply, I can send you some pictures directly and you can post them on your blog if you wish.
I was so weirded out when I saw this page! I just created something very similar as a school project. The main difference being that I used throw/catch events as a method for pattern detection. See poster #2 here: http://www.ece.cornell.edu/ece/academics/graduate/meng/meng-poster-session-2015.cfm
Please be kind and respectful to help make the comments section excellent. (Comment Policy)