Matthew- In addition to the specific recommendations provided by others about the role of stitching vias, I would recommend you follow my Rule #9: Never do a measurement or simulation without first anticipating what you expect to see. You can read a little more about this rule in the following two notes: http://www.bethesignal.net/bogatin/bts218-bogatins-rules-engineering-p-679.html?cPath0 http://www.designconcommunity.com/author.asp?section_id&07 <http://www.designconcommunity.com/author.asp?section_id&07&doc_id%2916> &doc_id%2916 If you have good engineering intuition about how the return currents flow when signals change return planes, you will more easily be able to evaluate the results from the simulation. You might also want to check out question #12 in this blog post: http://bethesignal.net/blog/?pE2 And toward the end of this webinar, https://www.bethesignal.net/bogatin/nma855-read-sparameters-like-book-p-684.html?cPath , posted for free viewing, I show an example of the measured insertion loss of a signal changing return planes, with no adjacent return via, and walk through a simple interpretation of how to think about the return currents propagating in the cavity mode of the planes. This will supplement what you learned from Yuriyâs fine paper. Note that when the diff signals are âin phaseâ itâs a common signal and will behave exactly the same as two single end signals in parallel when passing though the planes. --eric ******************************************************* Dr. Eric Bogatin, Signal Integrity Evangelist Bogatin Enterprises Setting the Standard for Signal Integrity Training web site: <http://www.bethesignal.com/> www.beTheSignal.com Blog: <http://www.bethesignal.com/blog> www.beTheSignal.com/blog Twitter @beTheSignal e: <mailto:eric@xxxxxxxxxxxxxxx> eric@xxxxxxxxxxxxxxx Laboratory: Connected Community Networks 105 S Sunset St, Suite J Longmont, CO 80501 USA cell: 913-424-4333 skype: eric.bogatin *********************************************** Msg: #5 in digest Subject: [SI-LIST] effects of stitching vias last attempt- without attachment From: Matthew Severini <Matthew.Severini@xxxxxxxxx> Date: Wed, 28 Nov 2012 13:08:52 -0500 Hello Experts, I would like to start out by saying thank you all for the wealth of information available on this list. I am about as green as it gets in this industry, and I have learned more from this list in the last few months than I thought possible. I have some results from a modeling project that I am involved in that my team simply cannot wrap our heads around. The purpose of the project is to investigate viability of both traditional and novel structures for high speed serial lines. One of the things we wanted to look at was effects of return path continuity. As an extreme case we built a few differential via models without, any ground stitching vias. The pictures below hopefully convey the stackups better than I can explain in words. The first one breaks half the return path, the second one totally breaks the return path, and the third one leaves the center plane totally floating. Each one has a stripline differential pair on the bottom signal plane coming into a pair of vias, which have a nicely tuned antipad, and transition up to a stripline differential pair on the top signal plane. Some care was taken to ensure the structure was pretty close to 100 ohm differential. In normal construction all of the ground planes would be tied together by a pair of PTHââ¬â¢s just outside this pair of vias, but we wanted to see how badly we could break the structure by not including these. The issue is that the signal integrity of this structure is not nearly as bad as we thought it would be. We can understand that if the pair is excited differentially, the signal can simply travel back the other wire, but when we excite them in phase with one another, how could it not just look like an open circuit? How are the return currents getting back? I (tried to attach) a plot of insertion loss in all cases. It is encouraging that there is some DC attenuation in the bottom two structures, but I would imagine it would be worse than it is. I understand that the return path is not totally broken for the (num_layers =2) case, but one would expect a lot of energy to be lost because any currents in the external planes are either lost, or must be radiated. The first structure we built was (num layers=2): G _______________________ S ____________ G _________ | __________ S ___________| G ______________________ Then (num layers=3): G _______________________ S ____________ G ________ | __________ S | G __________ | __________ S ___________| G ______________________ Then finally (num layers=4): G _______________________ S ____________ G _________ | __________ S | G __________ | __________ S | G __________ | __________ S ___________| G ______________________ Thank you so much for your help, I am trying to learn as much as I can, and any intuition I can gain on this issue would be extremely helpful. Thank you again, Matthew Severini Research & Development Engineering Endicott Interconnect Technologies, Inc. Phone: (607) 755-8119 Matthew Severini ------------------------------------------------------------------ To unsubscribe from si-list: si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field or to administer your membership from a web page, go to: //www.freelists.org/webpage/si-list For help: si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field List forum is accessible at: http://tech.groups.yahoo.com/group/si-list List archives are viewable at: //www.freelists.org/archives/si-list Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu