1 of 1 people found this helpful
With hyperlynx 8.2 you should be able to simulate that using the 3D full-wave field solver (if you have the correct licenses)
Could one of the guys from Mentor Graphics please confirm this? Is this the only solution, as we currently do not have a 3d EM license?
Although I am not a Mentor guy, I will try answering your question.
You are completely right, HL SI doesn’t properly model the signals return paths (it assumes the planes ideal) today. That is easy to verify by inspecting the Spice netlist that you can generate out of your topology.
HyperLynx 3D EM (previously called IE3D) is a new tool that Mentor has acquired last year from Zeland. The latest version 15.2 released at the beginning of this year has been partially integrated with HL SI v8.2. A full HL SI license will enable you to do 3D simulations of the vias (only) in LineSim. However if you want to export a routed board or a portion of it in HL 3D EM and perform 3D simulations in this environment you will have to purchase this tool.
Regarding Mentor’s plan to enhance HL SI capabilities in the area of signals’ return path modeling, from my understanding they are on the process of migrating ICX technology (field solvers) into HL SI. Consequently we hope that the next release (I guess v9.0) will provide this feature.
Maybe someone from Mentor could confirm and eventually elaborate.
2 of 2 people found this helpful
Crossing a split in a reference is a 3D phenomenon if you want to capture the full effect of the split (i.e. return current). The return current will be a combination of radiated fields as well as any bypassing in the reference planes. The only way to truely capture this behavior right now in HyperLynx would be with HyperLynx 3D EM, our fullwave 3D simulator.
In the next release of HyperLynx, we'll have support for capturing the impedance discontinuities seen due to the split in the reference plane, so it will be partial support for the impact of doing this.
Cheers everyone for the help. I had thought this for a while, however I wanted to be 100% sure.
Just to pick up this split plane simulation again now. I just tried to use the HyperLynx 3D EM Design System v15.30 to simulate this effect on a cut-out area exported directly from BoardSim v9.2. However, the first problem is that the differential impedance calculated by 3D EM for the differential pair far away from the split is already very wrong. Could you have a look into this? I can supply my example design if you want.