Because you want the panel to fit a non-square shape, with no room for "leftovers" - think making it round to fit in a doorknob hole or oval to go in place of a car brand badge, etc.

Theoretically you can. It's whether it all goes pear shaped when you take a Dremel to the board to make it

    OO    OOOO   OO
    OO    OOOO   OO
    OO    OOOO   OO
    O    OOOOOO   O
    O    OOOOOO   O
    O    OOOOOO   O
        OOOOOOOO
        OOOOOOOO

That... is a really good idea. I was thinking of cutting a flat piece of diffusing material to shape.

My intuition is that there would be pain in 3d printing a diffusion structure because slicers etc. would not be optimised for producing a homogenous solid. I would guess that 100% infill is actually something like 99.98% with tiny voids that stick out like a sore thumb when you shine a light through. I might be wrong about that, I'm not a 3d printing expert.

The principle of reshaping I think is awesome though. It might just be an issue of modulating brightness to counter any uneven distribution. It's got me thinking about a Faceted approach. 3d Print a faceted basin and then print a thin edge divider to sit in it. Fill it with something that sets solidly enough and makes a good diffuser (this too sounds like a war with bubbles). Take it out of the basin when set and you potentially have a nice faceted surface with each facet individually colourable.

Did you ever try to use LED's as an input device?

My next LED cube project will be oriented towards trying to do that.

Pretty cool way to fail

Not just accelerometer, but magnetometer too. :)

This means you can basically use the magicshifter3000 as a knob. Or, a slider.

Its very fun with MIDI.

:)

Donated to metalab through a member connection. Austria/Vienna has a bit of a 'recycle/re-use/repurpose' scene, and this was an early example of a metalab win:

https://metalab.at/

We devoured the panels and did all kinds of funky things with them.

https://metalab.at/wiki/Blinkofant/LED-Display_History

Yeah, expired a while ago, its a pretty old site and we have to get the old gang back together to fix it.

Thanks for checking out the magicShifter3000!

I've got a new one designed, currently in prototyping stages .. a bit more powerful and a lot more oriented towards audio/synthesis. Same form-factor, and probably MS3000 and midiShifter will have an ecosystem ..

Looks like they're WS2812B-2020, like on this PCB:

https://www.wemos.cc/en/latest/d1_mini_shield/8x8_rgb.html

I will make another small LED panel with ONE MORE LED!

Seems within reach of a hobbyist maker, at least the slightly unhinged kind (in the best meaning). Since considerable energy is needed but also avoiding excessive heating, integrating temperature measurement into the bulb might be prudent. Or at the very least having regulation on current+voltage. There might be risk of thermal runaway, if conductivity goes up when voltage is applied to the epoxy mixture.

I think the thing that fools us here is the fact that we're used to pixels being discrete, so it doesn't "look" low-res because the colors blend into each other. If you imagine that each LED produces a small circle onto the diffuser, and that this circle overlaps the ones around it a bit, it gets easier to see through the illusion.

It's much easier to understand when you can change the distance of the diffuser to the panel (which I did when testing), because then you can see the lights go from little squares with lots of dark space around them, to this, to big blobs of one color.

A diffuser is basically applying a convolution to the underlying pixel array (moving the diffuser away from the LED increases the convolution width, but unfortunately also reduces the intensity).

I don't think you can exploit that to show something higher resolution than the original pixel array,

Hm that's about 256 LEDs, must have been a 16x16 board. Sounds about right, good point.