I wanted to build my own totally mechnical flip-dot display for the bus stop machine, and after some experimenting I mostly figured out the mechanism I’ll be using. (This is in service of the bus stop interaction machine I’m building—if you’re unfamiliar take a look at the introductory post first to get caught up.)

First prototypes for mechanism

My first prototyping session ended with implementing three rectangular flipping pixels that flipped from the default position of dark (in this case, plain cardboard) to light (in this case, silver). A magnetic field of the correct orientation pushes on a magnet mounted to the pixel to induce the rotation. Here’s a gif of one pixel moving through its range of motion (imagine a magnet doing the pushing):

flipping pixel

It’s a bit difficult to tell from the image, but the axle is formed by passing a piece of straight steel wire through the corrugated path in the cardboard as shown a bit closer here:

corrugated swivel

I lined up three of these next to each other to experiment with slightly different positioning of the magnet. Using a steel wire as the axle wasn’t a great idea, because the magnets interacted with it.

three pixels

Second prototypes for mechanism refining, aesthetics, and machine reproducibility

I will be making the final version of this out of 1/8” acrylic (not cardboard) and that means I need to make some additional considerations, like 1) the axles won’t pass through the width of the board because it’s nigh impossible to drill laterally through 1/8” acrylic, and 2) the final design has to be laser-cuttable so I can make the hundreds of identical pixels I need.

To accommodate not being able to cut through the width of the plastic, I modified the design so that the axle sits on one surface, like so:

surface axle

In addition to moving the axle location, I wanted to imitate the circular pixels of the classic flip-dot display, so I started with hand cutting a more circular pixel, then making a laser-cuttable design that used my findings from the mechanical testing. Here’s a few different iterations next to each other:

pixel iterations

Assembly of the new roundish pixels went like this:

round pixel assembly

I made eight of them and built a small array. These are approximately 35mm wide and the eagle-eyed among you may have noticed that the top row doesn’t perfectly align with the bottom; I lasercut the two at different times and couldn’t line them up on the laser bed very well.

eight pixel array

Here’s what the eight pixel array looks like with silver coloring, indicating one of my favorite Tetris pieces:

Tetris array

You can’t tell from the still image, but the pixels really do responsively flip over quite nicely when even a weak opposing magnetic field passes behind them.

Playing with size

I have been prototyping in software to figure out how many pixels are really necessary to make an effective display for the purposes of the shared canvas. But I thought it would also be worthile to try lasercutting some really small pixels to see if they might work. So far, they don’t!

I scaled the plans down by a factor of about three, making the dots only about 10mm wide, and cut the backing from black acrylic and the dots from yellow. I started by making a big array in the backing but not so many dots, in case they didn’t work.

small plastic dots

The dots didn’t work! The thickness of the acrylic (1/8” or about 3mm) was now pretty significant compared with the overall size of the dots, and they couldn’t properly swivel any more.

Luckily, I found some scrap material made of some sort of paperboard with two differently colored sides. (Increasing the ease of finding just the right bit of scrap like this is exactly why I recently built shelves!) I used a piece of thread, glued at the centers of the dots, as my axle, figuring that the thread would have enough torsional give to permit the dots to freely rotate. I had to use tweezers and the smallest amount of hot glue I could apply to try to get these dots in the right spots.

gluing paper dots

I didn’t even finish the top row before I tried applying magnets to the dots, because if it wasn’t going to work I wanted to know that early on. And it wasn’t going to work! Notice how these dots are alternating:

alternating dots

That’s because at this scale, the magnets of the dots are relatively large and are interacting with their neighbors. Their lowest energy state is in alternating positions, so even if I push them into another cofiguration, as soon as I take my finger off, they return to this alternating form.

So given the mechanical difficulties I was having at that small scale, as well as the plainly much larger amount of work it would take to build so many more dots, I plan to stick to pixels at a scale at, or close to, the 35mm width I built in cardboard. (Of course, since I’m building a rectangular array, every linear decrease in pixel size leads to a quadratic increase in the number of pixels required to fill a given area.) The decision to stick with larger pixels is reinforced by my initial finding that people were able to have meaningful drawing interactions with each other using a 20×30 pixel array; since I’m aiming for a roughly 30”×18” drawing area, it would appear that these larger pixels are already in the right ballpark size-wise.

Up next

And miles to go before I sleep!

I need to:

  1. Work on prototyping at my chosen scale in acrylic instead of cardboard.
    • Play with the pivoting mount; perhaps glue straws onto the discs but there may be a better way
    • Investigate engraving the acrylic surface to provide a shallow trough for the axle to sit in
    • Tighten tolerances to match the flatness and uniformity of the material
  2. Design and test the mechanism that returns pixels to their dark state.
  3. Link these parts up with the plotter mechanism I’ve already built so that a magnet is driven by control knobs.