A couple of screenshots for those who can't wait;

The mirror and cylinder lenses are removed.  Five percent of the light is reflected. This is used to trigger the photo-diode.

For each ray drawn in freecad, the prism is in a different position. I draw them so that I know how to design the system.

A simulation using pyoptools at 42 degrees is shown below.

A ray goes through an ideal lens is refracted by a tilted prism at 42 degrees. Five percent of the light gets reflected at the second prism facet and is directed  to the diode as it hits the third facet at an angle greater than the critical angle (total reflection).

The prism is not fixed to anything at the moment. I still have to think about this. I envision two solution.
A metal base plane below the pcb using the magnetic force of the magnets or a mechanical push from above.

If the prism is assumed "perfect" and somehow sticks to the teflon outer ring, which does not rotate!!, the system is rotational invariant. 
A teflon cylinder is glued to the prism, at the bottom four magnets are placed which are 5 mm above the pcb.

Designing the thing will take some more time. An issue, I currently have is that there is too much play between the teflon inner and outer ring. Still, this might be easy to resolve. I might post a video when I send it to the fab.
I have also mostly replaced the raspberry pi with an ESP 32 board running micropython.
Most libraries have been ported to the board. Also the libraries which were in C or C++ (fomuflash and TCMStepper). I made a custom board, which can be put in the raspberry pi connector of my main board.