Ten Stacked Transparent OLEDs Make A Volumetric Display

Oct 11, 2026 Last reply: 3 hours ago 3 Replies
Jump to replies (3)

The hardest part here had to be the frame and the cabling, and it shows how much care went in. Several 3D-printed attempts failed before a working black frame came out, and fitting the fragile ribbon cables to each screen was described as a nightmare. Getting ten thin transparent panels lined up and wired without breaking anything is the kind of task that would stop most people, so seeing a finished handheld unit is satisfying.



The build itself is tidy. A custom board called TENEX V1, laid out in KiCad, carries an Adafruit Feather M4 and an MPU6050 for motion-based interaction, all running on 12 volts and programmed in the Arduino IDE. Solder paste went on through a metal stencil, parts were placed with tweezers and reflowed in a modified toaster oven, and the green PCB sits at the base under a stack of ten 128x56 transparent OLEDs from CrystalFontz, each with its own built-in polarizer and its own chip select line on SPI. The orange ribbon cable linking screen stack to board, plus the small silver screws holding it together, give it a proper product feel.



Depth comes from illuminating different layers of the transparent screens, and the demos make good use of it: rotating triangles, squares and circles, a stylized SH logo, and a particle simulation with blue points bouncing in 3D.



One trade-off stands out: those built-in polarizers block a lot of light, so brightness suffers. The maker calls it a proof of concept that can be scaled and thinks it can still be improved, which seems fair to me.



Project by SeanHodgins. Watch it on their YouTube channel:

formatting link


Giving every panel its own chip select line is the smart move here. It keeps all ten screens on one SPI bus instead of needing separate buses or extra controllers, which keeps the board simple. The trade-off is more pins used on the microcontroller, but with a Feather M4 there is room to spare, so that seems like the right call.

I would like to know how the ten panels are addressed in software. With one chip select per display on a shared SPI bus, is the frame data pushed to each layer in sequence, and does the Feather M4 keep up when the particle simulation runs across the full stack? Any refresh rate limits worth knowing about?

The gyro is sitting right there on the board, so a tilt-controlled ball rolling through a 3D maze would be a natural fit, with the maze walls split across the different layers. It would show off the motion sensing and the depth effect at the same time, and it feels like a small step beyond the bouncing particle demo.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required