@tony73 š I thought about I have to post news the next days how it's proceeding.
I went back to school in a figurative sense the last two months to learn more about robot kinematics, especially matrix calculations, so there was no development. I'll proceed with building soon.
Core elements are differential screws (I need them at multiple places), encoder support and diy harmonic drive, and finishing the RRF kinematics code. I'll use OpenCV for quality checks, it's a great program, I can recommend the book of Escriva/Laganiere about OpenCV 4.
I'll use OpenCV together with the cameras/lenses for
check the quality of the harmonic drive teeth
check printing result while printing
CNC position checking, 3D printing object position checking, SMD placement
homing position checking
measure precision of the kinematics (joints, steppers, belts, pulleys)
The patents of harmonic drives are interesting, I follow inventor Aubin with US5456139, EP0640778, US5662008 and similar from Ishikawa to optimize teeth. I try 3D printed parts first (using PP for the flexspline), but if it's not sufficient change to drilling with CNC with spring steel.