5 Replies Latest reply on May 31, 2016 11:27 PM by rocco_gruschwitz

    3D via modelling in hyperlynx v9.2


      I am using hyperlynx v9.2 for 3D EM simulation, i have exported 3D area of differential via from board sim and assigned the ports, while simulating in MGRID it asks for meshing frequency i gave it as 6GHz and then simulated upto 6GHz. In simulation window it asked for number of points i gave it as 100. It simulated within 2 minutes which is very less time as compared to other softwares. When i imported s parameters in touchstone viewer return loss(S11)  and insertion loss (S21) graphs are not proper.I think it is not simulating all points.

      Can you please help me where i am lagging.

        • 1. Re: 3D via modelling in hyperlynx v9.2

          hi, regarding the result of course I can't say a lot without seeing it. probably somewhere inside of the geometry a connection is not as it should be. But you are right, the solver doesn't solve all frequency points. It is able to reduce the number of required frequencies based on the already available results. BTW: have you tried the new 3d solver environment introduced with HL9.3? It is in many areas easier to handle and delivers quicker results than HL3D. just let me know if you need here more assistance.

          1 of 1 people found this helpful
          • 2. Re: 3D via modelling in hyperlynx v9.2

            Thanks for the reply.                                              I have used hyperlynx v9.3 nimbic it gave proper result, same via I am simulating in v9.2.I have exported in 3d then there is port creation wizard I assigned all four port as port with lower reference, then there is an option of number of segments I gave it as 12. While simulation can you tell me what other things I should take care of, and you mentioned some connections I might be missing, what type of connections, can you elaborate, because I assigned all the four ports.                                   And nimbic took 1hour to simulate same via, but v9.2 3D solver is creating S parameter within 5 minutes, surely I am lagging somewhere, please help me out of this.    

            • 3. Re: 3D via modelling in hyperlynx v9.2

              In HL3D I have seen very small opens between the polygon which you can look for with the check connection command. In 9.3 based on the HL Advanced Solver technology we don't see this effect. That's why I recommend this one. The difference in run time is rather large. Even if I don't know what machine or cluster you use I tend to say that 5minutes also with adaptive frequency sweep (reduction of frequencies) is too short when the same 3D area takes in HL 9.3 roughly 1 hour. Again, without seeing it is a bit difficult to judge. Just shoot me an email with the data and I'll have a look at it (rocco_gruschwitz@mentor.com).

              • 4. Re: 3D via modelling in hyperlynx v9.2


                I selected an option of adaptive intelli-fit in 3D solver, because of that it is simulating within 2 minutes, but while deselecting that option it takes about 30 minutes to simulate the entire results.

                • 5. Re: 3D via modelling in hyperlynx v9.2

                  That's what AIF was made for - total run time reduction be reducing the number of simulated frequencies and interpolation the frequencies between the simulated frequencies. You may see in the touchstone viewer that there are much more points in the model than simulated. If the result does not look as you would expect you need to identify the reason. Happy to help you doing this.