I need to know the orientation (rotation) of a turntable, in absolute degrees or radians. The turntable is not driven by a motor -- if it were I could use a stepper motor and use the motor counts to get the turntable position -- it's rotated by an external force (a human).
Fortunately, it rotates only through 180 (+/- 90) degrees; unfortunately, it's not possible to install a coaxial mechanical rot encoder. I can only work with the outer edge or the top surface. I've considered a friction drive on a mechanical encoder, and a belt drive; not really thrilled with those options. So... optical?
I have seen some downloadable gray-code disk encoder layouts, but they seem rather crude (resolution 45 degrees, that kind of thing). I was hoping for far more accuracy than that.
I thought of trying to repurpose the linear encoder strip from a junked inkjet printer, but have no idea how to go about it, what kind of sensor would be required etc. The samples I have seen look very finely graduated, so I'm not sure if my crude "hobby grade" optical sensors are suitable. Also, I would want to use reflectivity or albedo (can't easily get a light source behind the tape) and I'm not sure, but I think I recall that these print head position tapes use backlighting.
At present I am leaning away from optical encoding altogether, and towards a generic accelerometer/compass unit, but am not sure whether they are accurate enough for my purposes. Was wondering if anyone has experience with these and if so, would you recommend them (or not) for this type of application, i.e. absolute rotation encoder.
If anyone else has solved this problem and knows the best way to go at it (cheap, reliable, accurate), I'd be very interested to hear about your project.