|Phase 0 - Rounding the worm to 240 seconds an old bearing with the seal removed was marked off at approximate 0, 80, 160, 240 second intervals with white paint on the inner outer race and ball with 0 being the index.||Phase 1 - At the 80 second mark notice how a different ball is aligned with the top index mark of the outer race.|
|Phase 2 - At the 160 second mark notice again how a ball is aligned with the top index mark.||Phase 3 - At the 240 second mark a ball is once again aligned with the top index mark.|
There is definitely a correlation with the worm phase as it relates to the 76 second error (80 seconds with the new worm block)and physical position of the bearing race. This error exists even when using ceramic bearings so the material type doesn't matter. What both types of bearings have in common are the same physical sizes and both use 8 balls. The ceramics ball phase alignments are likely approximate which accounts for the 76 error encountered during testing. With the new block the 76 error was shifted to 80 seconds thereby eligible to be reduced using PEC to values well below 1 arc second putting it in the levels of seeing condition 'noise' and minimizing any impact on imaging with longer focal lengths.
So...getting back to why? The next logical conclusion is the axial load on the bearing is placing thrust on the inner race thereby causing the bearing race to ride up against the wall of the bearing grove and causing friction. These roller bearing types work better for radial loads whereby the balls ride in the center of the grove than axial loads.