2 Replies Latest reply on Oct 10, 2016 7:15 AM by vova2211

    Hyperlynx differential impedance calculation mistake



      I'm simulating an exported from Altium designer 16.1 Rigid Flex PCB board, which contains suppose to be 100 ohm diff pairs  . The Hyperlink calculates its Zdiff as 82 ohm for some reason .

      1. The designed PCB with correct PCB stack-up and Diff pare parameters where set in the Altium .

      2. Exported as HYP file from Altium and opened in Hyperlynx V9.4 .  All passed well, with correct traces drawing and layers stack-up . The ref / signal planes were selected .

      3. The selected diff traces where extracted from BoardSim to LineSim , in order to review a transmission lines topology and impedance.

      Somehow, the calculated by Hyperlynx differential embedded microistrip impedance is 82.1 ohm, while it suppose to be 100 ohm according to its layers stack-up and Er.

      Attachment 1.jpg - screen of LineSim with all the related  geometric parameters of transmission line with a wrong calculated Zdiff.

      Attachment 2.jpg - the correct calculated Zdiff by two known impedance calculators using the same trace and stack-up parameters.

      Where is my mistake ?

        • 1. Re: Hyperlynx differential impedance calculation mistake


          I have noticed that you selected, in Saturn PCB, ”Edge Coupled Ext” for the Differential Layer option. This would be valid if the differential pair was on an external layer.

          If you select ”Edge Cpld Int Asym” you get 80,497 Ohm, which is very close to what HyperLynx calculates.


          Saturn 6.89.png


          Best regards,


          Mircea Slanina

          1 of 1 people found this helpful
          • 2. Re: Hyperlynx differential impedance calculation mistake

            Dear fiend Mircea Slanina, thank you for your attention and response !

            You are right, my Saturn stackup model was wrong.

            I looked on this as on Microstip, because it is Flexible PCB where the flexible area actually has only 2 layers, and it is major area. So, it seems, Hyperlynx doesn't support Rigid Flexible PCBs with different stackups over the same PCB. Or, just the exporting process from Altium to Hyperlynx doesn't support it.


            But anyway, there is still significant mismatching between Saturn and Hypelynx diff impedance calculations .

            In case you are showing,  H2 layer is 5mil .

            In my case , the H2 of Saturn means the dielectric row2 (DE_Top_Layer)  in my stack-up, which is 25.2 mil .

            Here is correct Saturn model, and I still expecting from Hyperlynx to get the same result . So, as before, I still missing here anything.