I should've just bought a MIDI controller*.

That would be end of the story, but...

Instead, I designed a modular isomorphic system with Hall effect velocity sensing, per-key RGB LEDs per every single key, a custom binary configuration protocol, tons of firmware custom code pages, and a web app that lets you load different configuration (YES - microtonal too!) I regret nothing. I regret some things. Here's the whole story.

• Dad - can I have Lumatone?
• No, we have Lumatone at home...
• Lumatone at home:
  

Disclaimer - I have never touched Lumatone keyboard, I can only believe that it is a professional grade keyboard

That's how it begin, why isomorphic keyboards are so expensive? Now I know far more than I knew back in December 2025.

• Was it worth? Yes,
• Was it cheaper? Yes, hence, I though, it will be cheaper :) (thanks to our governments for the tariffs)
• Is it working? YES

Idea:
Build a fully modular MIDI controller with velocity awareness, after-touch. No compromises over number of modules, number of buttons we want to handle, full microtonal support possible. Do you want small MIDI module? No problem - 32 keys... or maybe you want the big one with 240 keys? No problem. Well, I've put a limit up to 512 keys, hence, it's kind of unrealistic anyone would find it useful (and yet that can be easily changed).

Limitations - we will use Gateron genty silent - unfortunately market is quite limited for kind of magnetic switches we can use or at least I was not able to find something with a bigger maximum travel. That's probably the biggest limitations and theoretically - easiest to address, as no other piece of hardware needs to be adjusted. Yes, internally we can re-calibrate and hall sensor are sensitive enough to measure the switch from centimeters away.

First step - refresh my rusty knowledge of designing electronic schematic - that was kind of a challenge, considering, I had only theoretical knowledge back from university time, but what for are all the data-sheets? And the fact that nowadays it is super easy to prototype? Nothing is gonna stop us! Well, maybe the fact, that with first three prototypes I was not super happy. From electronic point of view - all of them were working fine, but... no debug interface, not precise enough placement of keys, why did I use these simple RGB LEDs which are too small? Is using Ethernet connector really the best idea?