How to Build a Wi-Fi Connected Electromechanical 7-Segment Shadow Display

In this article, I’ll walk you through every step of how I built my very own Shadow Display—and show you how to create one yourself. I’ve chosen to turn mine into a sleek, minimalist clock, but the beauty of this open-source project lies in its flexibility: you can decide how many digits to display, separator style and adapt it in the way that suits your taste.

All the design files, firmware and assembly guides are freely available, so feel free to download, tinker and customize to your heart’s content.

If you don’t have the time—or just can’t be bothered to build that display yourself—swing by my store. You might find something there that catches your eye.

Electromagnets 101: When current flows through a wire coil, it generates a magnetic field. One end of the coil becomes a north pole and the other a south pole (depending on current direction). If you place a permanent magnet nearby, that magnet will experience a force: opposite poles attract and like poles repel. In our display, each segment has a small neodymium magnet attached. By energizing a coil behind the segment, we can either attract or repel the magnet, causing the segment to flip inward or outward.

Magnetic Force Considerations: The strength of the magnetic push/pull depends on a few factors: the coil’s current, the number of turns in the coil, and whether there’s a core material to concentrate the field. Higher current and more turns make a stronger electromagnet. Traditional electromagnets use an iron core to greatly boost the field, but our PCB coils are air-core (just copper and air). To compensate, I designed the PCB coils with multiple turns and layers to generate as strong a field as possible within the space constraints. Through some experimentation, I found that about 30 turns spread over 4 PCB layers worked well for flipping the small magnet/segment reliably,

In summary, by leveraging magnetic attraction and repulsion, we can get a crisp mechanical motion for each segment. Next, let’s see how we drive these coils electrically – this is where the H-bridge drivers come in.

To flip a segment, we need to send current through its coil in either direction as needed. Manually swapping wires to reverse a coil would be impractical, so we use an electronic solution: an H-bridge circuit. An H-bridge is a configuration of four switches (often transistors or MOSFETs) arranged like the letter “H”, with the load (our coil) in the center. By closing two opposite switches at a time, current can flow through the coil in one direction; closing the other two switches sends current the opposite way. This is a common method to control DC motors as well – it allows you to drive motors forward or backward, or in our case, magnetize the coil as north-up or south-up.

Each H-bridge needs two control lines. So if you’re working with, say, six modules—that’s 42 segments, meaning you’d need 84 GPIOs. You’d also have to route all those signals between modules, which would be an absolute nightmare. Not to mention, the display wouldn’t be expandable.
That’s where shift registers (e.g. 74HC595) come in.

Here’s how it works in our context: The microcontroller sends a stream of bits (corresponding to the desired on/off states of each segment coil) into the shift registers serially. Once all bits are shifted in, a latch pin is toggled to transfer those bits to the output pins of the 74HC595s, which are connected to the H-bridge driver inputs. By daisy-chaining, for example, two 74HC595s, we get 16 outputs.

Each segment needs two outputs to drive its H-bridge—so that’d be 14 outputs per module. To keep the wiring simple, I just stick two 8-bit shift registers in each module. Yeah, it “wastes” a couple of pins, but those chips are dirt cheap, so who cares? Plus I use one of those outputs to feed a feedback signal back to the microcontroller. So basically, there’s really only one output that isn’t used.

Main Board

On the main board, aside from the microcontroller, there’s also an RTC chip and a coin battery. So if you want to use it as a standalone clock, completely offline—you totally can.
As for the power connector—on my setup, it’s on the main module, but you can place it on any module. That means your display can be powered from the right side, the left, or even from the middle—whatever works best for your layout.

Extension Board

The PCB for the extension modules is identical—the only difference is what gets soldered onto it.
It’s actually way simpler. It really just uses shift registers, H-bridges, and an LDO to power the electronics.

In the original design, active segments appeared white, while inactive ones blended into the black enclosure. In this updated version, everything is white. That immediately raises an important question: how do you distinguish between active and inactive segments? The answer lies not in color, but in light—specifically, shadow.

Inactive segments sit perfectly flush with the front panel. When activated, a segment rises a few millimeters, creating a subtle shadow that reveals its shape. It’s a minimal, elegant effect driven entirely by depth and lighting.