After spending a few days pondering, admiring, and salivating over this, I decided to start sketching. Whatever came to mine, I wrote it down and tried to draw. I’m talking about the Everbright video that was posted not too long ago. Here it is again in case you need a refresher: http://www.hero-design.com/everbright/
First question that popped up on the original post was whether that was a potentiometer or an encoder. My guess tells me it’s an encoder. Pots have an upper and lower limit, they only turn so far. And while this is by itself not a show stopper, if you notice in the video, as some point, the whole board gets reset back to black. If those were pots, that’s a lot of motors to spin them back to their start position. An encoder is much better for this. Specifically, an incrementing encoder. An absolute encoder will remember its position when you power cycle it, whereas an incrementing encoder just starts pulsing, it doesn’t care where it is.
Next, the mechanical part, what’s in it, what spins, what doesn’t. And this is where all we can do is speculate and be as creative as we want to be. But, there are a couple of things that are obvious:
a) an encoder, which means sensor and encoder wheel
b) light source, and in this case I’m going with LEDs
c) the outer shell rotated, but it’s attached to a backing, so that tells me there’s a stationary part to it, which helps because of the encoder
d) the lens … ah the magic lens. In the video it’s black, but not just black, it’s a matte black, not shiny. The problem with using black acrylic is that you will affect the light shining through, there is no way you will get the same vibrant colors as shown.
Rather than trying to figure out what it was, I went a different direction and opted for frosted white instead. I have some samples which I tested and it made me realize something else about using LEDs: edge shine. Rather than putting LEDs directly below the acrylic shining through it, why not put them so they shine on the edge. If you take a piece of clear acrylic, put some scratches on it, and shine a light along an edge, those scratches will “light up”. So taking a piece of frosted acrylic and shining a light along its edge made it light up nice and uniform. Aha moment!
So I went with all of that. What you see below is a possible setup, and it’s designed to be put together (as opposed to have pre-molded pieces), and for those who want to, also print the pieces to be assembled.
Going through the whole piece, you have:
- a bottom plate (black) that remains stationary - this can be used to attach the whole thing to something. There’s a hole in the center to allow for wires to come through
- on top of that, a ball bearing (red) with an inner diameter that fits snugly onto the center hole of the bottom plate
- on top of that comes the bottom of the rotating piece (yellow), it fits snugly on the outer diameter of the ball bearing so it can rotate freely from the stationary plate
- an encoder wheel (orange) sits on the rotating plate
- then you get a “cup” of sorts (white) which holds the LEDs in it, and the optical sensor on the bottom
- the “cup” is affixed to the bottom, stationary plate with a retaining plug (blue) which also has a hole all the way through for wires
- and finally the outer shell itself (black), which also holds the frosted acrylic, which clips onto the bottom rotating plate
You can watch the video to see how it all comes together. The outer shell, with acrylic lens, rotates via the bottom plate that’s attached to the outside of the bearing. On the inside, the encoder wheel is attached to the rotating plate so that it also rotates, however everything else inside is stationary, including the bottom plate.
From the sideways view, you can see the alignment of the LEDs versus the edge of the acrylic. It’s not necessarily ideal, but it works. I can probably lower it even more so it sits at the right height, but I wanted a retaining ring that the lens sits in, and I’m also toying with the idea that one could also stick a piece of black acrylic under it, making it twice as thick.
In the top wireframe you can see where the sensor sits (after I rotated it 45 degrees so it’s not right ontop of an LED, oops) and how the encoder wheel passes through it. You can also see where the LEDs are.
While the drawing I made has 4 LEDs in it, I’m thinking one could probably get away with just 3, which also makes it easier because you can use a WS2801 IC and wire it so it drives all three LEDs (the same way their cheap strips are made.) It makes things easier that way. Why didn’t I just use a strip? Because they’re too dang wide at 12~14mm. Right now the PCB holding the LED is only 7mm “tall”. Route all the wiring for both the LED IC as well as the sensor through the center hole and you have one compact little unit that’s 2.8" in diameter and about 1.1" tall. Not too shabby to be honest.
There’s a little bit of room still under the “cup” to add another thin PCB with the IC that drives the LEDs on it, or the design can be slightly modified to accommodate that. Or one could go nuts and stick APA102s in the thing and control each one separately for even more psychedelic colors. It just complicates things later though. I’d rather address the whole pod as “one” addressable unit, regardless of how many LEDs are in it as opposed to having to do calculations for x-amount of LEDs in each one.
Remember, this is just one of many possible ways those things work, I just came up with what I thought was feasible to recreate. So much so that I may, I just may, try to actually build one. Though it’ll be a while before I tackle it as I have a 40+ hours video project I’m working on till well into November.
There you have it. Comment, brain storm, throw out your own ideas. There are different ways to light up anything … let’s hear it!
