Using a motor as a rotary encoder is obviously not new. You can read the previous log entry to see how I was inspired.

But every attempt I saw so far used the motor (BLDC, PMSM or hybrid stepper) as a generator, and measured the phase voltage looking for a transition edge or keeping track of amplified levels. The main problem with this is, if the rotation speed is too low, the output level is also low and movement is not detected. To counter that, amplifiers are often used. But that won't solve the problem completely and at one point introduces a new one: noise.

My approach is different. I use the coils in the motor as coupled inductors and the rotor changes the coupling ratio. This isn't a new thing either, just google "inductive encoder". Even more, I'm pretty sure I'm not the first person to try this, but I didn't find a widely public writing. Meanwhile, I saw the reactions to the previous projects, many liked them despite the drawbacks.

That's why I do this, I try to give a better solution.