Sonnet questions about EM simulations
I am using Sonnet for my EM simulations. I have to simulate asymmetric uniform coupled lines and use the results in some calculations preferably done by hand.
I used +1 and +1 port definition for my c mode and +1 and -1 port definition for Pi mode. Propagation constants should follow from results immediatly. I wonder how to get c and Pi mode characteristic impedances from EM simulations (should be 4 of them, 3 are independant, I am getting 2 by my method of simulation-port definitios).
Furthermore, these are not enough to characterize asymmetrically coupled lines, voltage ratio Rc or RPI (at least one) is needed to be able to recover Z or Y matrix. How to get this form EM simulatons?
I am thinking of simulating of 0 length line (reference-calibration planes for two sides of the box meeting at the same point) or quarterwavelength, get Z or Y matrix from Sonnet directly and work these parameters out from matrices and known theoretical expressions.
I need this data as initial point to verify some theoretical results when some connections to this 4 port are made.
flyhigh
Hi Flyhigh -- For asymmetric coupled lines, even and odd modes do not exist. This is because using even (+1,+1) or odd (+1,-1) excitations do not yield PMC (Perfect Magnetic Conductor) or PEC (Perfect Electric Conductor) walls. Judging from the way you phrase your question, I think you are already aware of this, I mention this here for the benefit of others.
For asymmetric lines, there are multiple ways to view the modes. As far as I know, there is no standard way in any of the frameworks to describe an asymmetric coupled line in terms of its modes.
Thus, probably the best way would be, similiar to what you suggest, de-embed a length of line, not to zero length, but to one cell long. In the Sonnet data viewer (from xgeom do Project->View Response->Add to Graph), you can use Output->N Coupled Line Model to output the R, L, C, and G matrices for your very short length of line. You should be able to calculate your desired information from that. If not, you can also output Y or Z parameters from the same Output menu.
If your line (before de-embedding) is too short, your result might not be good. To be sure you are OK, repeat your analysis with a different length line, but still de-embedded to one cell long. You should get essentially the same answer. As a rule-of-thumb, before de-embedding, your line should be longer than about two substrate thicknesses or line widths.
Questions like this are also be nice to see on the Sonnet forum, www.sonnetsoftware.com, click on the right side "Forum".