I have a fairly basic question about trace simulation. How do you model a trace that is sandwiched between two 3V3 planes but is driven from logic supplied by another plane (1V8 for example)? For simulation is it realistic to treat the 3V3 plane as Ground? Especially if the 3V3 plane is fairly close to an adjacent ground plane? Regards,Bryan Ackerly.
bryan 5 years 10 months 28 days
The best answer
You can select the best answer for current question!
Answered bybradb 5 years 10 months 28 days
Hi Bryan,To judge the signal integrity of your interconnect you will need to model the trace as well as its complete return path. For the portion of trace in the isolated cavity formed by the 3V3 power rail it will locally have return currents in those two planes. The issue is how that return current flows from the 1V8-referenced source location (and termination location) to those regions of the 3V3 rail nets. This concept is often called "power-aware SI". You can simulate this complete behavior with many simulators, not necessarily 3D, as long as you define the geometry with full representation of the return path and your simulator supports such geometry.Best regards, -BradP.S. As mentioned by others, please be careful with your simulation boundary conditions - both explicit and implicit. As a geeky detail, you can get into trouble with a so-called 2.5D Method of Moments simulation if you use power planes of infinite lateral extent by representing the voids with equivalent magnetic currents. In such case, there is an implicit DC path between all the planes so you will get results different than you measure for the physically finite sized planes. This DC connection can be explained in multiple ways: e.g. a short at "infinity" or an infinite capacitance between planes. In such simulation you should apply finite extent (and therefore numerically meshed) planes.-----
Answered byshlepnev 5 years 10 months 28 days
Hi Bryan,The trace simulation should be done with all surrounding planes treated asthe reference conductors. A TEM wave propagating along the trace will havereturn currents on the planes, no matter what those planes are. Though, thisis only a part of the problem in this case. Another, and probably moreimportant, part is how that TEM wave is formed or how you design thetransition to the trace. Such transition cannot be localized in your casefor single-ended link and for the common mode in differential link. This isbecause of the natural reference conductors cannot be connected to providethe path for return currents. Non-localized structures can be simulatedbelow microwave frequency band with a solver that includes transmissionplane models (to simulate distributed return paths). A 3D EM analysis isrequired at microwave and millimeter wave frequencies, but it may be uselessin the case of non-localized problem. The analysis of non-localizedstructures is not useful in isolation from the rest of the board with anyboundary conditions - change of simulation area or boundary conditionschanges the analysis outcome in that case. Best regards,YuriyYuriy Shlepnev, Ph.D.President, Simberian Inc.3030 S Torrey Pines Dr. Las Vegas, NV 89146, USAOffice +1-702-876-2882; Fax +1-702-482-7903Cell +1-206-409-2368; Virtual +1-408-627-7706Skype: shlepnevwww.simberian.com Simbeor - Accurate, Fast, Easy and Affordable Electromagnetic SignalIntegrity Software2010 and 2011 DesignVision Award Winner-----