0 Replies Latest reply on May 25, 2017 3:57 AM by jorge.alvarez

    PDN analysis issue with S-Parameter models

    jorge.alvarez

      Working with HyperLynx PI, I have found an issue when I tried to make a PDN analysis of my design making use of S-Parameter models for the capacitors.

       

      To reduce the explanation, I'm going to reduce the circuit to a single capacitor, but actually I facing the problem in an anlysis of the complete PCB's PDN.

       

      Then, analyzing a single 470nF ceramic capacitor, I get the right response, no matter whether I work with RLC or S-Parameter model.

      You can see it at below picture.

      ceramic_470nF_linesim.good.01.JPG

      • Purple. RLC model. Impedance vs. Frequency, without inter-plane capacitance.
      • Red. RLC model. Impedance vs. Frequency, with inter-plane capacitance.

       

      The problem comes when I change the RLC model to S-Parameter model.

      The behaviour shown at Touchstone viewer is not the right response for the capacitor.

       

      Here, is a comparison between RLC and S-Parameter model for the same capacitor.

      ceramic_470nF_linesim.wrong.01.JPG

      • Purple. RLC model. Impedance vs. Frequency, without inter-plane capacitance.
      • Red. RLC model. Impedance vs. Frequency, with inter-plane capacitance.
      • Orange. S2P model. Impedance vs. Frequency, without inter-plane capacitance.
      • Yellow. S2P model. Impedance vs. Frequency, with inter-plane capacitance.

       

      Is the root cause in the S2P model? Or maybe is it needed to include any input to HyperLynx?

       

      * The S-Parameter models are obtained from manufacturer's website (http://ksim.kemet.com/Default.aspx), and before work with them, I have checked the right behaviour of them.

       

      Any help will be welcomed in order to understand the problem, and of course, find a solution.

       

      BR!