In Part 1, I tried an AMOLED touchscreen dial. The screen was excellent, but fast turns missed steps, so reliable adjustments required a moderate turning speed. With around 30 physical clicks per turn, fine adjustments also required several full rotations to cover the range. I started looking for a dial with a precise magnetic encoder and found something even more interesting: a knob where the steps themselves are programmable.

Scott Bezek’s amazing SmartKnob project

Like many other hardware enthusiasts, I was fascinated by Scott Bezek’s revolutionary SmartKnob project, which combines a magnetic encoder with a motor.

A magnetic angle encoder measures the orientation of a magnet attached to the knob. It reads the absolute angle within a revolution without mechanical contacts, rather than relying only on counting pulses. Scott’s short demo of a magnetic encoder knob shows this part of the concept.

Scott’s SmartKnob adds a motor that applies resistance according to the measured angle. This creates virtual clicks, on/off snaps and end stops, with spacing and strength defined in software. I recommend watching the SmartKnob videos linked in his repository to see these behaviours in action.

Unfortunately, my preliminary search suggested the parts alone would cost over $300. The assembled SeedLabs SmartKnob DevKit was quoted at approximately $412 plus $27 shipping to Denmark.

A cheaper SmartKnob alternative: X-Knob

I then found the X-Knob project and bought an assembled unit on AliExpress for approximately $71 including delivery. It used the same magnetic encoder and motor concept, but also offered a built-in battery, an LVGL-based X-TRACK interface and Wi-Fi/MQTT control for Home Assistant. Those software features were still listed as future plans in Scott’s original README.

X-Knob also supports Bluetooth Surface Dial functionality for a PC, seven knob modes combining detents, bounds and end stops, and vibration feedback when pressed.

The device arrived with several demos installed. It was less polished than the AMOLED dial and felt more like a proof of concept, which was what I expected. Trying the on/off snaps and virtual end stops immediately confirmed why I had wanted one. Even before adapting it for my own use, I could recommend it as a fun device for hardware enthusiasts interested in haptics.

The video below shows the built-in demos alongside my own 100-step configuration with stronger haptic feedback (turn on sound). I disabled the RGB lights around the base because I thought they made the device look cheap.

Here is what the X-knob internals look like:

What the magnetic encoder and motor improve

The magnetic encoder measures the knob’s angle, while the motor creates steps with programmable spacing and strength. This lets the physical rotation, tactile steps and on-screen indicator follow the same scale, as the built-in demos already showed. I simply adjusted the firmware to suit my percentage range and preferred feedback strength.

The video below shows the physical rotation and on-screen indicator staying in sync during both slow and fast turns. This was exactly what I had been missing with the AMOLED dial: precise adjustments and reliable tracking without needing several full rotations to cover the range.

Limitations of the X-Knob

One limitation of the X-Knob is that pressing it produces mushy, unsatisfying haptic feedback. There is virtually no travel, considerable force is needed, and presses did not always register reliably in my setup. I much preferred the physical click of the AMOLED dial. (The original SmartKnob or the more expensive SeedLabs version may feel different)

The unit also lacks touch controls and separate physical buttons. Its LCD is small, with lower resolution and contrast than the AMOLED screen, leaving less room for controls and context.

On a smooth surface, the whole unit rotates as I turn against the motor’s resistance, which is not ideal. It would need a much heavier base or a grippier underside to stay in place, and the device is not exactly lightweight to begin with.

The built-in battery also needs recharging after less than a day of use.

(A minor gripe: the steps feel good during normal turning, but holding the knob between them produces a faint, silent vibration as the motor resists my hand. The resulting tingling sensation feels unnatural and breaks the illusion of mechanical clicks.)

Finally, the ergonomic limitations from Part 1 remain: reading the display means looking down at my hand and adjusting my grip and wrist position to keep it visible. I cannot simply hold the remote in whatever way feels most comfortable.

Pros

• Programmable haptic steps, on/off snaps and end stops.
• Adjustable step spacing and strength, so physical rotation can match the display.
• Precise magnetic angle sensing that handled fast turns and fine adjustments well in my setup.
• Assembled hardware with demos, Bluetooth and Wi-Fi/MQTT support.
• Built-in battery and an affordable price.
• Great toy for experimenting with haptics

Cons

• Mushy press feedback that requires considerable force and does not always register reliably.
• No touch or separate physical buttons, and a small LCD with limited resolution and contrast.
• Tall form factor and a base that can rotate on smooth surfaces under the motor’s resistance.
• Less than a day of battery life in my use.
• Reading the display means looking down and adjusting my grip and wrist position.
• Holding the knob between steps reveals motor vibrations rather than a natural mechanical feel.

Unfortunately, the link in my AliExpress order history no longer works, but the Makerfabs MaTouch SmartKnob, listed at $69, looks like similar alternative. It is a different product that I haven’t tested, but it even includes a touchscreen, which the X-Knob does not have.

The X-Knob showed me how much a magnetic encoder and programmable steps could improve rotation. But it also helped clarify what I wanted from a sofa remote: a higher-quality feel, reliable physical buttons I could use without looking, and fine adjustments that also worked during fast turns. I wanted flexible controls without relying on a touchscreen.

In Part 3, I revisit those requirements and consider whether a separate display would make more sense than putting the screen inside the remote.