Many thanks, Steve, Dave and Ege There seems to be some agreement, at least, on the theory and the difficulty of presentation. I have made enquiries with two major vendors but not found any offer of a good solution. Some points on package modelling: 1. One of the papers by Steer's team at NCSU explains the idea of local reference nodes for simulation of a microwave circuit. I suggest this has immediate applications for SI and IC package modelling because it addresses the problem of a distributed microwave or RF network, such as a package or a connector, together with organising the connection nodes for interface to an integrated circuit simulator. The NCSU team are working on fREEDA: http://guppie.egrc.ncsu.edu/freeda/ Also, Leeds Univ in UK is developing a unified approach, led by Christopher Snowden. However the *full solution* for a distributed packaged IC from DC to THz is not yet commercially available and is unlikely to be so for at least a couple of years. 2. The package alone can be accurately modelled using a fullwave 3D tool such as Ansoft HFSS or CST Microwave Studio (which is my own choice so far). Both have travelling wave ports and discrete nodal ports. The travelling wave ports seem to be unsuitable for a structure with several terminals, as there is a requirement for them to be well separated for accuracy. 3. The discrete ports have limited accuracy because of their finite size. However, they can be made to work if carefully set up. Each port has its own separate signal and return terminal. MWS has Discrete Ports and HFSS has Gap Ports. 4. Because of the lack of the *full solution* of a 3D non-linear full-wave simulator incorporating characterised semiconductor behaviour, we have to divide the system into two parts. The first half is the chip, simulated in Spice or similar using a nodal schematic picture, which is fundamentally flawed of course, because wave couplings are not accurately represented, even if transmission lines are used for the interconnects. However, the chip is small enough that we can live with the errors, and we have no choice. An important point is that the schematic has a single common ground terminal to which all power and signal voltages are referred - this is part of the lie. The second half is the package and local PCB interconnects, represented as accurately as we can in a 3D EM simulator. 5. The EM simulator produces a file of s parameters. This is defined for n ports. In the 3D model, these n ports had 2n terminals but this information is not implicitly carried along with the s-parameter file. 6. We now leave the EM simulator behind and consider the circuit simulator. How can these s-parameters be used with the chip circuit? This is where Steer's local reference node idea comes in. These can be used in the n-port element within the Cadence circuit simulator, together with the circuit simulation of the chip. I must say I don't have a taste for modified nodal admittance matrices, and economic pressures prevent their detailed consideration. But perhaps we can gain some insight, remembering we need to concentrate on true-as-possible representations of currents and differential voltages. Dave's simple example of a microstrip with a gap port might help the understanding here. 7. There is another difficulty, which Ege Engin has just pointed out. The PCB side of the package has ports referred to the PCB ground plane - I see no problem in joining those reference terminals together in the circuit simulator. That is what we normally do when we consider a network of transmission lines with spatially separated ports on a common ground conductor. But the IC interface of the package has a number of bond pad terminals quite close together. The ports of the package model would most likely be referred to a common paddle connection, but this may not always be possible. If there is a paddle connecting the IC substrate it works like a local common ground plane, partly screening the IC from wave interaction between its terminals and with the rest of the package. If this approximation is valid, we can then join the IC reference terminals together to form a single "local reference node", representing the paddle connection. 8. If we have faith in the validity of the s-parameters, and because the IC ports and their local reference node have no interaction with the PCB side, we can now join the paddle connection to common ground. It remains to be added that the s-parameters must be valid from DC to RF, so that the total effects of DC resistance, internal inductance, RF resistance and all AC and wave couplings are included. This may require stitching s-parameter results together from different frequency band runs in the EM simulator. _________________________________________________________ Zetex Semiconductors - Solutions for an analog world EID Award Winners for 'Best Use of Technology' 2003 for the AcoustarTM ZXCW8100 End-to-End Digital Audio Amplifier Controller http://www.zetex.com _________________________________________________________ ###################################################################### E-MAILS are susceptible to interference. You should not assume that the contents originated from the sender or the Zetex Group or that they have been accurately reproduced from their original form. Zetex accepts no responsibility for information, errors or omissions in this e-mail nor for its use or misuse nor for any act committed or omitted in connection with this communication. If in doubt, please verify the authenticity with the sender. ###################################################################### ------------------------------------------------------------------ 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.org 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