I am performing an analysis on a simple shaft supported by 2 bearings. I first performed the analysis on only the shaft model using bearing fixtures. I created split faces on the shaft where the bearings were located and where the loads were applied. The bearing fixtures are both self-aligning and one bearing fixture prevents shaft rotation. I put mesh a fine mesh control in the corner radii of the shaft where I expect high stresses (5 elements across corner radius). My stresses and deflection seemed reasonable. Also the animated deflection appeared to be correct in how the part deforms. When I tried to get resultant forces on the bearings, it showed all zeros.
I setup the analysis again as an assembly with simple bearings and the shaft. I used bearing connectors this time, both self-aligning and one preventing shaft rotation. I also setup a no penetration contact set between the shaft and bearings. I fixed the bottom of the bearing housings. Again the animated displacement appears correct and the part is deforming properly. However, the stress in the shaft is roughly 1/3 as much as when I ran the analysis with bearing fixtures.
The deformation is very similar but slightly different in the 2 studies as well. When I used bearing connectors, the bearings do deflect and the shaft pulls away from the bearing race just slightly as expected.
I've attached a picture showing the 2 results and setup. Thank you for your help!