PDN resonance mode excited by high speed signal

Hi experts,I'm trying to simulate the phenomena of PDN resonance mode getting excited by the EM field generated from high speed signal. Here is my modeling:1.      A 2''x4'' 6-layer PCB with the following stackup:      TOP       SIG      PL02      GND      PL03      PWR      PL04      PWR      PL05      GND      BOT       SIG2.      Route one trace from one end of the board to the other end passing through 4 via from TOP to BOT. the reason of inserting vias in the trace is that as of now most hybrid simulators if not all of them, neglect the coupling btw trace and plane. Meanwhile the coupling btw via and plane is much more significant and is accounted by the solvers (Thanks, Yuriy and Brad).3.      The excitation connected to one end of the trace is a pulse of 1V ampl, 20ps tr/tf, 1ns high time, and 3ns period.  Also I hooked one 1V DC source btw the power and the ground planes though it may not be necessary. Both sources are located at one end of the board, and at the other ends two resistor of 1Ohm are placed as the loads to each source respectively.4.      I ran the transient sim then convert the resultant waveform of the voltage across PDN into freq. spectrum. But looks the none of the spectrum components meet any of the PDN resonance freq. I tried to lower the load to 0.01Ohm, still didn't see the excitation occurring.5.      The sim tool used is Cadence Speed2000.I will check my setup again tomorrow. Please advice if you see anything wrong in above. Thanks.Best Regards,Sherman ChenSignal IntegrityEMC Global Hardware EngineeringTel: +86 21 60951100-3329
sherman.chen 5 years 9 months 2 days

2 answers


The best answer


You can select the best answer for current question!
Answered byistvan.novak 5 years 9 months 2 days
You can look at (simulate) the impedance profile of the power plane structure at the locations of the vias and use directly the repetition frequency for the signal excitation that hits an impedance peak.Regards,Istvan NovakOracleOn 12/22/2014 1:20 PM, Chen, Sherman wrote:> Hi experts,> I'm trying to simulate the phenomena of PDN resonance mode getting excited by > the EM field generated from high speed signal. Here is my modeling:> 1.      A 2''x4'' 6-layer PCB with the following stackup:>        TOP       SIG>        PL02      GND>        PL03      PWR>        PL04      PWR>        PL05      GND>        BOT       SIG>> 2.      Route one trace from one end of the board to the other end passing > through 4 via from TOP to BOT. the reason of inserting vias in the trace is > that as of now most hybrid simulators if not all of them, neglect the > coupling btw trace and plane. Meanwhile the coupling btw via and plane is > much more significant and is accounted by the solvers (Thanks, Yuriy and > Brad).> 3.      The excitation connected to one end of the trace is a pulse of 1V > ampl, 20ps tr/tf, 1ns high time, and 3ns period.  Also I hooked one 1V DC > source btw the power and the ground planes though it may not be necessary. > Both sources are located at one end of the board, and at the other ends two > resistor of 1Ohm are placed as the loads to each source respectively.> 4.      I ran the transient sim then convert the resultant waveform of the > voltage across PDN into freq. spectrum. But looks the none of the spectrum > components meet any of the PDN resonance freq. I tried to lower the load to > 0.01Ohm, still didn't see the excitation occurring.> 5.      The sim tool used is Cadence Speed2000.>> I will check my setup again tomorrow. Please advice if you see anything wrong > in above. Thanks.>> Best Regards,>> Sherman Chen> Signal Integrity> EMC Global Hardware Engineering> Tel: +86 21 60951100-3329>>>
Answered byscott 5 years 9 months 2 days
ShermanGiven the difference in impedance of the PDN and the signal trace, youmight not see anything.  But, if you place two signal traces and theirassociated vias penetrating through the PDN, you should definitely see thePDN peaking show up as crosstalk peaking. You are better off simulatingthese sorts of things with a Frequency doman solver where you see therichness of the spectrum.  To get that with a time domain solver requiresthat you use a very long run length PRBS pattern.ScottScott McMorrowConsultant - R&D16 Stormy Brook RdFalmouth, ME 04105(401) 284-1827 Businesshttp://www.teraspeed.comOn Mon, Dec 22, 2014 at 1:20 PM, Chen, Sherman  wrote:> Hi experts,> I'm trying to simulate the phenomena of PDN resonance mode getting excited> by the EM field generated from high speed signal. Here is my modeling:> 1.      A 2''x4'' 6-layer PCB with the following stackup:>       TOP       SIG>       PL02      GND>       PL03      PWR>       PL04      PWR>       PL05      GND>       BOT       SIG>> 2.      Route one trace from one end of the board to the other end passing> through 4 via from TOP to BOT. the reason of inserting vias in the trace is> that as of now most hybrid simulators if not all of them, neglect the> coupling btw trace and plane. Meanwhile the coupling btw via and plane is> much more significant and is accounted by the solvers (Thanks, Yuriy and> Brad).> 3.      The excitation connected to one end of the trace is a pulse of 1V> ampl, 20ps tr/tf, 1ns high time, and 3ns period.  Also I hooked one 1V DC> source btw the power and the ground planes though it may not be necessary.> Both sources are located at one end of the board, and at the other ends two> resistor of 1Ohm are placed as the loads to each source respectively.> 4.      I ran the transient sim then convert the resultant waveform of the> voltage across PDN into freq. spectrum. But looks the none of the spectrum> components meet any of the PDN resonance freq. I tried to lower the load to> 0.01Ohm, still didn't see the excitation occurring.> 5.      The sim tool used is Cadence Speed2000.>> I will check my setup again tomorrow. Please advice if you see anything> wrong in above. Thanks.>> Best Regards,>> Sherman Chen> Signal Integrity> EMC Global Hardware Engineering> Tel: +86 21 60951100-3329>>>>