11 Replies Latest reply on Jan 31, 2011 2:54 PM by Chris Nelson

    Tracking down why a study fails to solve/wierd issue

    Greg McFadden

      Can someone  suggest any method to figure out how to fix a problem I have been having.  for some reason, once in a while a study will fail to solve.  I immediately get the"failed to solve" error.  it is somewhat random, but after that happens, any study much more complex than a cube will fail. The only way to get any study to run successfully is to create a cube (or something similar) and run a study. sooner or later it will run successfully, and after that, all the other studies will also run successfully.  restarting/reboooting does not do anything.

       

      sw2010 sp4

       

      thanks
      Greg

        • Re: Tracking down why a study fails to solve/wierd issue
          Ryan Werner

          Hi Greg,

           

          That is a strange problem.  Do you happen to be running SWX 2010?  I have noticed that since I upgraded to 2010 the Iterative solver is very flaky about what models it will run.  Strangely enough I have noticed that it matters how I open the model I am trying to analyze.  For instance, if I open a file by double clicking on it in Windows Explorer the Iterative Solver will fail immediately.  If I then open the same file using the SWX Open function then the Iterative will work.  Luckily, I have found that no matter how I open a file the Direct Sparse solver will always work.  So, if I want to use the Iterative Solver and it fails right away I will try reopening it through SWX and it usually works.  If it still does not work than I will try the Direct Sparse.

           

          Hopefully this is helpful, good luck.

           

          Ryan W.

            • Re: Tracking down why a study fails to solve/wierd issue
              Greg McFadden

              It is truly strange....  on further examination, the only way to push past the errors is to create a new cube, run a stupidly simple  simulation on it, and then the rest of the studies will run....  A saved simple cube will fail.   Every single time I open simulation, I have to create a new file, run a stupidly simple simulation, and then the rest will work.  very very strange.  I am not sure what the acronym SWX stands for, we are running solidworks office premium 2010 x64 edition patched to SP4.0

            • Re: Tracking down why a study fails to solve/wierd issue
              Tony Gatta

              I've had the same exact problem for some time now.  The iterative solver, as mentioned,

              instantly fails.  going to Direct solver works fine (every time) -- but i believe makes the solution

              more computationally intensive.

               

              -Tony

              • Re: Tracking down why a study fails to solve/wierd issue
                Shawn Miller

                I am also having this issue.  The Automatic or Iterative solver fails (even on a simple cylinder with 1 flat).  The direct solver seems to work.  I am also running SW 2010 Premium SP 3.1.  Thanks for the advice about the Direct solver.  I thought I was going nuts.

                 

                Shawn M.

                  • Re: Tracking down why a study fails to solve/wierd issue
                    Damien Murphy

                    I have also come accross the issue, running a HP z400, Direct Sparse is fine but the FFEPlus fails even on very simple model.  Previously similar problems were caused by the HP ProtectTools:  Under Embedded Security State, the Chip needed to be disabled.  However this z400 does not have the same security manager to check it.  Has anybody had experience on changing these settings.

                    Damien

                      • Re: Tracking down why a study fails to solve/wierd issue
                        Alex Fletcher

                        We are experincing the same problem, with SW 2011 Sp 1.

                        Currently once a study fails, we go though older studies until one runs successfully.

                        Then the original study will run fine. We will try switching to the direct solver next time we have this issue.

                         

                        Has anyone raised it with their VAR yet?

                         

                        BR

                         

                        Alex

                          • Re: Tracking down why a study fails to solve/wierd issue
                            Greg McFadden

                            I have not.  Generally when it happens I go straight to a simple cube I have set up as a direct solver simulation and back to work.  I've never been able to record the issue.

                              • Re: Tracking down why a study fails to solve/wierd issue
                                Damien Murphy

                                I have some susggestions below that came from SolidWorks:

                                 

                                1) Reduce to minimum “User Account Control” on this machine; the option is present under Control Panel->User Accounts. A restart of the machine is required after doing that.
                                2) Check the behaviour in an administrative login.
                                3) Check if any other security tools are active on this machine.

                                4) Go to the properties of the study and point the Result folder to a new location where in the user has read/write permission.
                                Probably create a new folder and point it to that location.
                                5) Or, go to Simulation->Options->Default Options-> use the option “User defined” and set a folder over there let it have read/write permission.
                                After doing that create a new study, copy all the loads and boundary conditions and then check the behaviour.
                                6) Create a new User login with administrative privileges or even a normal user and then check the behaviour in it.

                                 

                                Damien

                              • Re: Tracking down why a study fails to solve/wierd issue
                                Ryan Werner

                                I did mention it to my VAR and he had no explanation for the problem beyond giving the typical advice.  It was while I was trying to capture the problem with Rx that I noticed that if the Iterative failed and I closed the model and reopened through SolidWorks, not Windows Explorer, that the Iterative would then work.  Go figure.  I have no idea what difference that would make but it works everytime.

                                 

                                Ryan W.

                          • Re: Tracking down why a study fails to solve/wierd issue
                            Chris Nelson

                            For what it's worth, I'm having the same problem.  SW2011, sp1, 64 bit Window 7 machine.  If I log in as the administrator, the problem always goes away, but I'm looking for a better solution.

                             

                            -Chris