thanks sir fr this video , i know this video has no relation with CST , but ADS has relation with S11 so if you can help me or to suggest TO me some materials has clear explanation for this extractions from CST RF sensor design could you please show me how can i extract an equivalent circuit from CST design to draw it in ADS .. ?
Another fantastic video. The mismatch between the measured and simulated data can be due to several reasons: one one hand, the model used to obtain the simulated data is an approximation of the reality. Maybe the frequency span is also too large. Models tend to well behave only around an operating point. Finally, if the optimization problem is strongly non-convex, as it probably is, heuristic based optimization methods can, sometimes, be stuck in local minima. Twiking the optimization algorithm is a pain and, sometimes, can be harder than solving the original problem :)
Yes heuristics need to be tailored for the particular optimization problem, this standard differential evolution method for another highly nonlinear problem in another discipline. The model in this case is a generalized transmission line model, but one method would be to reduce some of the variables assuming you know them. So for example if this were measurements of an actual transmission line and you knew the equivalent circuit mode for the input buffer within the reliable frequency range, you could reduce the number of parameters by entering this load impedance directly into the model. Another candidate for a transmission line would be the R value since you can get it directly from the trace geometry and it will always hold. Roughness is a really tough one though because you now have to assume a standard roughness model or you need to have tabulated data for roughness, lots of research has gone into this.
Thank you, the method I use in the video is very quick just to illustrate the method and the goals. There is a good reason for doing this. It is used so that you can take a complex system, like a transmission line or connector, and use it in a standard circuit simulation. You could also take multiples of these extracted models and connect them together, then you can do SPICE simulations or cascaded linear network simulations.
Thanks for your efforts and inspiring videos. It will be awesome if you explain differential evolution also.
Noted
thank you sir, that's an awesome explanation, can you post a video of doing the extraction in advanced design system.
Nice. Will check the sheet. Good work, Sir.
Enjoy!
thanks sir fr this video ,
i know this video has no relation with CST , but ADS has relation with S11 so if you can help me or to suggest TO me some materials has clear explanation for this extractions from CST RF sensor design
could you please show me how can i extract an equivalent circuit from CST design to draw it in ADS .. ?
Sir, please put video on extraction complex permittivity an permeability from S- parameters
Another fantastic video. The mismatch between the measured and simulated data can be due to several reasons: one one hand, the model used to obtain the simulated data is an approximation of the reality. Maybe the frequency span is also too large. Models tend to well behave only around an operating point. Finally, if the optimization problem is strongly non-convex, as it probably is, heuristic based optimization methods can, sometimes, be stuck in local minima. Twiking the optimization algorithm is a pain and, sometimes, can be harder than solving the original problem :)
Yes heuristics need to be tailored for the particular optimization problem, this standard differential evolution method for another highly nonlinear problem in another discipline. The model in this case is a generalized transmission line model, but one method would be to reduce some of the variables assuming you know them. So for example if this were measurements of an actual transmission line and you knew the equivalent circuit mode for the input buffer within the reliable frequency range, you could reduce the number of parameters by entering this load impedance directly into the model. Another candidate for a transmission line would be the R value since you can get it directly from the trace geometry and it will always hold. Roughness is a really tough one though because you now have to assume a standard roughness model or you need to have tabulated data for roughness, lots of research has gone into this.
Interesting!!! It's like using inverse problem technique !!!!
Thank you, the method I use in the video is very quick just to illustrate the method and the goals. There is a good reason for doing this. It is used so that you can take a complex system, like a transmission line or connector, and use it in a standard circuit simulation. You could also take multiples of these extracted models and connect them together, then you can do SPICE simulations or cascaded linear network simulations.
Thank you, it works perfect!
You're welcome!
i am the opposite lol