Hi,
I have an S parameter file generated by EMX for my oscillator. It has 56 ports and was simulated from DC-900GHz with a 1GHz step. When I used bbspice as the interpolation method for the nport, the rational fitting takes almost 1 day and still has not finished. Could you suggest some techniques I could use to reduce the fitting time? Here I have kept the nport compression as default. I'm curious since Tawna mentioned in an earlier thread that she used to simulate an n-port with 405 ports (https://community.cadence.com/cadence_technology_forums/f/rf-design/37979/how-to-compare-what-spectre-interprets-from-the-raw-s-parameter-file/1352837#1352837).
I'm using Spectre Version 16.1.0.510.isr10 64bit -- 13 Oct 2017, Virtuoso IC6.1.7-64b.500.15, EMX 5.12, interface date 15/7/2020.
I include here the settings for my nport, and the screenshot from transient simulation showing the iterations of the fitting process.
Many thanks and regards,
Anh
