The Case of the Missing Ground: An EM Mystery

Feb 1, 2017 | Posted by: Dr. John Dunn

A colleague approached me recently with a vexing electromagnetic (EM) simulation problem.  An NI AWR Design Environment customer had been working on a module design for acoustic filters and needed to understand the RF signal path.  A simplified version of the layout—two chips connected to a carrier package by bond wires drawn in Analyst™ 3D finite-element method EM simulator—is shown in the figure to the right.

My colleague expected the insertion loss to be one and the return loss to be zero.  After all, we are simply connecting Ports 1 and 2 together with bond wires and line sections.  (The chips are simply through lines.)  But, instead the return loss was 6.6 dB at low frequency!  Why?  Yes, the bond wires were connected properly.  An open circuit would have an infinite return loss, not 6.6 dB.

Based on the sleuthing work I did with my colleague to solve this EM mystery, I have since written a new application example titled, “5 Tips for Successful EM Simulation.” This document highlights the five best practices I use to solve EM issues for a problem of this type and describes how I worked through each step in order to solve the mystery. 

Nope, I’m not going to give away the answer.  You have to read the example.

The five tips I use when working with EM simulation tools are:

  • Simplify the geometry whenever and wherever you can.
  • Check that your S-parameters make physical sense by running three simple tests: connectivity, passivity, and energy conservation.
  • Understand the wave ports and the EM modes they use.
  • Trace the return path of your currents when analyzing the S-parameters.
  • Understand the boundary conditions at the edges of the geometry.

In the example, you’ll see how I use these concepts to unearth the mysterious behavior of the circuit.  I hope you enjoy discovering the hidden secrets of the missing ground as much as I did. (Oops, did I just give the secret away?)