It's kinda ridiculous that seemingly three quarters of the people who write graycode decoder firmware get it wrong, despite the correct methods being publicly available with sample code since 15-20 years at least.

It's also an assignment that close to every electrical/computer engineer would have had to do as part of their degree.

I guess you only need to understand ~50% of the course material to get your degree though.

Sounds like clicky 3-pin A/B bi-phase type though. I’ve done stupid things and tried to read that and two buttons with single ADC. That jumped around when Vref fluctuates.

An "optical rotary encoder"? I doubt it. I bet it's a mechanical rotary encoder. And that it randomly jumps means their code for reading the encoder sucks. Something like that should not happen if properly implemented. I will take a rotary encoder over a potentiometer every time.

Potentiometers have that problem too when they get worn, which they always do if they also double as on/off switch like on most cheap radios.

They could use something like an AS5048A for 14-bit (effectively analog to humans) magnetic position sensing instead of a potentiometer or variable inductor.

Considering they already need digital control circuitry to interface to a touch screen it shouldn't be significantly cost-prohibitive to use that to make knobs that are immune to oxidization.

Kinda costly? I wonder if a dead simple 10kΩ B-curve potentiometer would do, though they drift a lot

Just about every analog lighting board I've ever seen has huge dead zones on the end of the linear pots on some channels.

I'm surprised the potentiometers work at all given the abuse they go through. Early desks had no memory, so it was often a huge rush to set the next scene in the fader bank before it was needed.

Yes, I keep a can of contact cleaner around to get rid of the oxidisation.

until it wears off or catches something between the slider and the base