11 Replies Latest reply on Dec 18, 2014 3:16 AM by Jared Conway

    Bearing Connector versus Bearing Fixture -- 1/3 stress value?

    Kyle Huck

      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!

       

      Shaft Results.png

        • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
          Jared Conway

          what is the loading condition? just gravity/weight?

           

          have you scoured the forum on bearing connectors and restraints? a lot of caveats to be careful about.

           

          ignore stresses to start, what are the displacements doing, you said they are close, but how close.

           

          how are the displacement results fixing the bearing races instead of using bearing connectors?

            • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
              Kyle Huck

              Jared, thanks for the feedback.  The loading condition is 5561N downward at each end of the shaft (outside of each bearing, symmetric loading).  Additionally, at one end of the shaft, there is 12820N load at a downward angle on the inside of the bearing.  You can see the Force arrows in the picture I attached in my original post, sorry that wasn't clear in my first post.

               

              When I stated the displacements were close, I wasn't specific enough.  I meant the displacement form or deformation was similar in shape.  The magnitude of deflection is quite different however (which is likely why the stresses are so different).  In the model where I used the shaft only and bearing fixtures, the displacement is 0.147 inches at the maximum point (you can see in the photo in my first post,  slightly left of center of the shaft towards the end with 2 forces).

               

              In the assembly using bearing connectors, the maximum displacement is at a similar location but is only 0.015 inches.

               

              I have looked through many posts and online/Solidworks help on bearing fixtures and bearing connectors.  It's possible I'm still overlooking something but that's why I'm here 

               

              I'm not sure I understand your last question about how the displacement results are fixing the bearing races.  In the model where I used bearing fixtures, I made split faces on the shaft where the bearings are located and made them self-aligning and one preventing shaft rotation.  In the assembly, I set the bearing connectors up the same way and also created a no-penetration contact set between the bearing race and the shaft.  I fixed the bottom face of the 2 bolt flange on the bearing housings.

               

              Again, the shape of the deformation is similar but in the study using bearing fixtures, since the displacement was much higher, the angle of the shaft at the bearing location is greater than in the assembly model using bearing connectors (See picture attached below).  The displacement values at the bearing is similar in both studies (about .005 inches in both studies).  As you move away from the bearing, the displacement increases more rapidly in the bearing fixture model since it is at a steeper angle coming out of the bearing due to the greater overall displacement/deformation of the shaft.

               

              One difference is in the bearing at the end with one load.  In the bearing connector model, it seems the shaft is at a much shallower angle.  In other words, that bearing doesn't have to move as far to "self-align" with the shaft deformation.  In the model with bearing fixtures, the shaft deformation is slightly more parabolic so both bearings are tipped at similar angles to "self-align".  I tried to show this in the picture below.

               

              Thank you again for your help!

              BEARING DISPLACEMENT.png

              DISPLACEMENT ANGLE.png

                • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                  Jared Conway

                  try running the analysis without the self alignment and without the prevent rotation enabled, let me know if it fails

                   

                  I think your problem is that you're getting something similar to what would happen if you had soft springs enabled and unfortunately that is the reason why bearing connectors and restraints cause people issues.

                    • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                      Kyle Huck

                      Jared, I turned off the self alignment and rotation prevention.  In the assembly model using bearing connectors, the analysis ran ok and the max displacement reduced slightly from .015" to .007" likely due to the bearings holding the shaft in position at the ends thus reducing its displacement.

                       

                      I also did this on the shaft part analysis with bearing fixtures.  The analysis failed and says "Model may not have adequate fixtures."

                       

                      Then I ran the shaft part study again and left the bearing fixture self alignment turned off, but I set one bearing fixture to prevent rotation.  The study ran ok.  Now the maximum displacement is only 0.001" which is closer to the 0.007" in the assembly study.  The displacement of 0.001" occurs at the shaft ends and also near the center of the shaft.  In the shaft assembly study, the shaft ends also show approximately 0.001" displacement but the maximum discplacement of 0.007" occurs near the center of the shaft.

                       

                      So what does this all mean?  In reality, I have self aligning bearings.  The displacements are very similar in the two models if I run them without self aligning bearings.  And although the displacements are similar, won't the non-aligning bearings increase the stress in the shaft near the bearing?  I'm not sure which result to have faith in.

                       

                      Thank you again for your help!

                        • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                          Jared Conway

                          have you looked at the kb article about what the bearing connector does on the back end? i'd recommend taking a look at that to start

                           

                          when you say you aren't sure which one you should have confidence in, i'm not sure what you're referring to as i'm not following what your testing resulted in.

                            • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                              Kyle Huck

                              I haven't seen the article, do you have a link?

                               

                              I'm still confused as to why I had to turn the self-alignment off on the bearing fixtures to get similar displacements as the assembly setup with bearing connectors.  I've made a simple table showing my results below.

                               

                                   Restraint                      Self-Aligning?               Prevent Rotation                       Max Displacement

                                Bearing Fixture                       Yes                                      Yes                                                0.147"

                                Bearing Fixture                        No                                       No                                Failed (inadequate restraints)

                                Bearing Fixture                        No                                       Yes                                                0.001"              

                              Bearing Connectors                 Yes                                      Yes                                                0.015"    

                              Bearing Connectors                  No                                       No                                                 0.007" 

                               

                              The most similar displacement results are having self-alignment turned off resulting in displacements of 0.007" and 0.001".  However, in reality I do have self-aligning bearings.  The results of the analyses vary drastically when I have the self-alignment turned on.  The result with bearing connectors doesn't change too much with self-alignment turned on but the "bearing fixtures" setup changes dramatically.

                                • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                                  Jared Conway

                                  bearing article is in the solidworks kb, customerportal.solidworks.com, you'll have to login with your account to get access to it. some of it is in the helpfile. self aligning links to a node in the center, without self-aligning uses the flat face of the bearing.

                                   

                                  i am suspect of your results for bearing connector no, no. so i don't think you can actually compare with your bearing fixture no, yes. they are both different setups. and anything with prevent rotation on, again i am skeptical of as it is similar to enabling soft springs. again great articles in the kb about how it will affect your results.

                                   

                                  my suggestion still holds that you have a case that you can't use a bearing connector. why does your assembly need to be restrained or connected using a bearing connector? what are you gaining from having it?

                                    • Re: Bearing Connector versus Bearing Fixture -- 1/3 stress value?
                                      Kyle Huck

                                      Jared, thanks for the reply.  I am also skeptical of the No/No Bearing connector results.  Unfortunately, when I turn off prevent rotation with the bearing fixture model, it fails and says there are inadequate restraints.

                                       

                                      I initially performed the analysis with just the shaft part and used bearing fixtures.  The reason I tried using the bearing connectors in an assembly, is my initial analysis with just the shaft and bearing fixtures gave me zeroes for all my bearing reaction loads.  I wanted to see the reaction loads at the bearings and the results would list them but all of the values were zero.  I thought this was maybe because they were "fixtures" and acted as "fixed".  Any ideas on that?

                                       

                                      When I set it up using bearing connectors as parts in an assembly, I did get resultant bearing loads in my results but that's when I noticed my displacement and stress values were drastically different from my initial model so I questioned it thinking that both setups with the same loads should've given similar results.

                                       

                                      Lastly, pardon my ignorance, what is kb?  I'll try logging into the customer portal to find the article.  Thank you again for your help.