Scott, I agree the microstrip structure and equations exist always when a trace is routed above a plane. However, the current distribution may sometimes change the characteristic impedance. For example when a trace is routed across a split power plane the characteristic impedance increases due to the discontinuity in the return current path. Cosmin Iorga NoiseCoupling.com --- On Wed, 1/28/09, Scott McMorrow <scott@xxxxxxxxxxxxx> wrote: From: Scott McMorrow <scott@xxxxxxxxxxxxx> Subject: Re: [SI-LIST] Re: BGA vias outside the package To: "Cosmin Iorga" <ci249534@xxxxxxxxx> Cc: si-list@xxxxxxxxxxxxx, "V S" <for_si2003@xxxxxxxxx> Date: Wednesday, January 28, 2009, 2:21 PM Cosmin With all due respect, a microstrip formula for power traces sure can be used. As long as the microstrip power and ground traces are above a plane, they are transmission lines and can be decomposed as such. This is no different than performing a PEEC decomposition. There will be transition issues at the boundaries between the trace and the parallel plates of the power system, but at these frequencies they are minimal. Will there be error? Yes, but as a first approximation, the microstrip traces will add incremental inductance (and therefore incremental noise) to the power system. If you want to worst case the design, then the microstrip analysis can be performed with the reference plane maximally far away from the trace (say the thickness of the board). I am, of course, assuming that the remainder of the power delivery network is modeled for via, spreading inductance, and capacitor inductance, using well-known analytical methods, as described in many papers. http://home.att.net/~istvan.novak/papers.html http://www.teraspeed.com/papers/cap_considerations_fpga_pds.pdf http://www.teraspeed.com/papers/DC08_FullWaveCapacitorModeling_paper.pdf http://www.teraspeed.com/papers/DC08_FullWaveCapacitorModeling_pres.pdf http://www.teraspeed.com/papers/TF7_Bypass%20capacitor_inductance.pdf http://www.teraspeed.com/papers/TF7_Bypass%20capacitor_inductance.pdf http://www.teraspeed.com/papers/stack_up_vias_pdn_public.pdf regards, Scott Scott McMorrow Teraspeed Consulting Group LLC 121 North River Drive Narragansett, RI 02882 (401) 284-1827 Business (401) 284-1840 Fax http://www.teraspeed.com Teraspeed® is the registered service mark of Teraspeed Consulting Group LLC Cosmin Iorga wrote: I don't think the microstrip formula can be used for the 1.8V and 3.3V traces, since the return current does not flow through the underneath plane. The return current flows through the ground trace, so I suggest that an evaluation of the loop inductance would give more accurate results than the microstrip analysis. Cosmin Iorga, NoiseCoupling.com --- On Wed, 1/28/09, V S <for_si2003@xxxxxxxxx> wrote: From: V S <for_si2003@xxxxxxxxx> Subject: [SI-LIST] Re: BGA vias outside the package To: si-list@xxxxxxxxxxxxx Date: Wednesday, January 28, 2009, 1:17 PM Scott, First I will complement your answer by adding - 1. Your have 4 pins on 1.8V rail. Assuming the four traces coming from the four balls have no coupling the total inductance will be equivalent to the parallel combination of the inductance. 2. Find the change in current dI expected in your 1.8V rail. IF dV is the maximum change in the voltage that your 1.8V rail can tolerate then the maximum tolerable impedance Z is given by dV/dI. 3. Make sure that the impedance across all the frequencies as calculated in the Scott's suggestion is less than this Z. You will need to repeat the things for your 3.3V rail. For calculating dI, take into account the maximum current consumption and the take into account current due to charging the all IO pins simultaneously. If C is the capacitive load, and if there are N IO pins, then this switching current can be gives as NCdV/dT... You should be able to find C on the datasheet. dV/dT can also be found on the datasheet. N is the maximum number of the IO pins. This switching current should be applicable to 3.3V supply. I am not sure about 1.8V. I believe, 1.8V is only the core voltage with not IO. This will give you a pass or fail answer - or a ball park number. If you can achieve the result with you existing design that is good. Otherwise, increase the number of capacitors, use reduced series inductance capacitors. Recalculate the check if you are meeting the required impedance target. Vikas Shukla --- On Wed, 1/28/09, Scott McMorrow <scott@xxxxxxxxxxxxx> wrote: From: Scott McMorrow <scott@xxxxxxxxxxxxx> Subject: [SI-LIST] Re: BGA vias outside the package To: Cc: si-list@xxxxxxxxxxxxx Date: Wednesday, January 28, 2009, 12:43 PM Ryan Use a 2D field solver to compute the impedance of the trace you'll use to connect the outer layer planes to the balls. If microstrip, prop delay is approximately 150 ps/in. Once you know your trace impedance you can compute the incremental inductance. Z = sqrt(L/C) tpd = sqrt(LC) solving for Inductance L = Z x tpd For 50 ohm microstrip that would be 7.5 nH of inductance per linear inch of trace. For 40 ohm microstrip = 6 nH/inch For 30 ohm microstrip = 4.nH/inch For 20 ohm microstrip = 3 nH/inch You can use this, along with available publicly published analytical formulas for planes and vias to compute the total inductance in the power delivery system up to your ball pads, to determine you power system impedance vs. frequency. regards, Scott Scott McMorrow Teraspeed Consulting Group LLC 121 North River Drive Narragansett, RI 02882 (401) 284-1827 Business (401) 284-1840 Fax http://www.teraspeed.com Teraspeed® is the registered service mark of Teraspeed Consulting Group LLC Ryan Sequeira wrote: Would like to understand the effects of supplying power and ground to the BGA by connecting vias to the VCC/GND plains and traces outside the BGA package. Pros: Allows for larger vias outside the package. Cons: Compromised noise performance through trace length added to VCC/GND trace length. Is there something else that needs to be taken care of.... The BGA device is a CPLD, IO=3.3V, Core=1.8V, Fmax > 180MHz, VCCO(3.3V) - 11 pins, VCC(1.8V) - 4 pins, GND - 15 pins The device is a 132csBGA, 0.5mm pitch, 0.3mm pad. So cannot afford to drop the vias within the BGA area. Microvias would be too expensive... Ryan ______________________________________________________________________________ DISCLAIMER The information contained in this e-mail message and/or attachments to it may contain confidential or privileged information. If you are not the intended recipient, any dissemination, use, review, distribution, printing or copying of the information contained in this e-mail message and/or attachments to it are strictly prohibited. If you have received this communication in error, please notify us by reply e-mail or directly to netsupport@xxxxxxxxxx or telephone and immediately and permanently delete the message and any attachments. Thank you. ______________________________________________________________________________ This email has been scrubbed for your protection by SecureMX. For more information visit http://securemx.in ______________________________________________________________________________ ------------------------------------------------------------------ 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 technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ 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 technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ 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 technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ 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 technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu ------------------------------------------------------------------ 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 technical documents are available at: http://www.si-list.net List archives are viewable at: //www.freelists.org/archives/si-list or at our remote archives: http://groups.yahoo.com/group/si-list/messages Old (prior to June 6, 2001) list archives are viewable at: http://www.qsl.net/wb6tpu