From the Agilent datasheet: Input capacitance1 Cm = 0.09 pF Cm is between tips Cg = 0.26 pF Cg is to ground for each tip Cdiff = 0.21 pF Differential mode capacitance = Cm + Cg/2 Cse = 0.35 pF Single-ended mode capacitance = Cm + Cg Tom Dagostino Teraspeed(R) Labs 13610 SW Harness Lane Beaverton, OR 97008 503-430-1065 503-430-1285 FAX tom@xxxxxxxxxxxxx www.teraspeed.com Teraspeed Consulting Group LLC 121 North River Drive Narragansett, RI 02882 401-284-1827 From: Bob Tarasewicz [mailto:Bob.Tarasewicz@xxxxxxxxxx] Sent: Friday, October 03, 2008 12:11 PM To: tom@xxxxxxxxxxxxx Cc: 'Cotton, Chris'; 'Istvan Novak - Board Design Technology'; si-list@xxxxxxxxxxxxx Subject: RE: [SI-LIST] Re: IBIS simulation LVPECL Tom, I am using Agilent InifinMax 1132A 5GHz differential probe with a E2677A solder-in probe head. I don't know how to model that probe/head, but I tried adding a fair amount of C and/or R at the receiver input pads of the simultation but couldn't simulate the same "step" that I see in the lab. Right now my best guess is that the package model isn't accurate, but not sure. It is a mult-core uP in a 45x45mm PBGA package with 1mm pitch. thanks, Bob Tarasewicz "Tom Dagostino" <tom@xxxxxxxxxxxxx> 10/03/2008 01:22 PM Please respond to <tom@xxxxxxxxxxxxx> To "'Bob Tarasewicz'" <Bob.Tarasewicz@xxxxxxxxxx>, "'Istvan Novak - Board Design Technology'" <Istvan.Novak@xxxxxxx>, "'Cotton, Chris'" <Chris.Cotton@xxxxxxxxxxx> cc <si-list@xxxxxxxxxxxxx> Subject RE: [SI-LIST] Re: IBIS simulation LVPECL What you are measuring and what you are simulating are not the same. When you put the probe on the net you are adding an additional set of parasitics from the probe that are not in the net you are simulating. And from what you describe below the half pF you changed the package model by is likely about the magnitude of the probe's parasitics. Try adding the probe to the simulation and see what you get. Tom Dagostino Teraspeed(R) Labs 13610 SW Harness Lane Beaverton, OR 97008 503-430-1065 503-430-1285 FAX tom@xxxxxxxxxxxxx www.teraspeed.com Teraspeed Consulting Group LLC 121 North River Drive Narragansett, RI 02882 401-284-1827 -----Original Message----- From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx] On Behalf Of Bob Tarasewicz Sent: Friday, October 03, 2008 9:02 AM To: Istvan Novak - Board Design Technology; Cotton, Chris Cc: si-list@xxxxxxxxxxxxx Subject: [SI-LIST] Re: IBIS simulation LVPECL Istvan & Chris The step is about 200ps wide. I am measuring at vias right under the BGA pads. I don't know how accurate the package model is; the processor vendor provided it. Here is the package model info for the 2 reciever pins AV21 and AU21 AV21 Len=0 L=0.1n R=1.7m / Len=11391.99 L=0.4p C=0.11f R=0.06m / Len=0 R=3m / Len=0 L=0.4n R=2.54m / Len=0 R=3m / Len=0 L=0.08n R=0.04m C= 0.03p / AU21 Len=0 L=0.1n R=1.7m / Len=10354.47 L=0.4p C=0.11f R=0.06m / Len=0 R=3m / Len=0 L=0.4n R=2.54m / Len=0 R=3m / Len=0 L=0.08n R=0.04m C= 0.03p / I have to make some pretty significant changes in the receiver ibis model 'Define Package Model' in order for simulation to start showing a "step". If I change the 'Len' by a factor of 2x and the C by a factor of 5x (ie make Len=20000 and C=0.55f for both AV21 and AU21) , then simultion will show a "step". However, I'm not sure if making these kinds of changes to Len and C is still within a reasonable range or not. Thanks for your responses. Bob Istvan Novak - Board Design Technology <Istvan.Novak@xxxxxxx> Sent by: Istvan.Novak@xxxxxxx 10/03/2008 09:56 AM To Bob Tarasewicz <Bob.Tarasewicz@xxxxxxxxxx> cc si-list@xxxxxxxxxxxxx Subject Re: [SI-LIST] IBIS simulation LVPECL Bob, How wide is the step? Could it come from the round-trip delay from the pin connection where you measure, to the die, including possible via(s), package traces, bond wires -if any - ? Regards, Istvan On 10/03/08 10:42, Bob Tarasewicz wrote: > My simulations of an LVPECL net (apprx 3" long, 'Y' terminated just beyond > receiver) do not seem to be correlating at all with in-lab measurements. > In the past I've had very good luck with simulations agreeing pretty well > with actual scope measurements; however, this is my first experience > simulating LVPECL. Is there a fundamental reason why perhaps IBIS models > won't model the behavior of LVPECL accurately? Or is it that the faster > edge rate of LVPECL is highlighting flaws/limitiations in my simulation > methodolgy and/or tool? > Simulations "at the pin" as well as "at the die" show nice rising/falling > edges at the receiver. However, scope captures with a good diff probe > show a 'step' in the rising/falling edge at the receiver. The 'step' > looks like what one would expect halfway down a series-terminated net; not > sure why I'm seeing a 'step' with an end-terminated net. > > Any guidance would be apprecitated. > > Bob Tarasewicz > > > ------------------------------------------------------------------ > 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