what solver are you using? what is the operating frequency and what is the mesh size?

I think I solved the above issue by running a stationary study of the circuit first to get the initial values.

I have a related problem though that is quite general so hopefully should be easy to answer.

I'm linking the circuit to my 3d model via lumped ports on my model, but I also have ports on my 3d model for the propagation of a wave, is it wrong to have 2 types of ports? No matter what I do to my circuit (components I have in it, parameters I change) it doesn't seem to effect my s parameters from my frequency domain study.

Any ideas off the top of your head why this might be?

How it is supposed to change? How are you setting up the split ring? material properties? are you using an specific module?

I'm trying to use the Nicholson-Ross-Weir retrieval method (with the Smith adaptation for negative index metamaterials) to simulate wave propagation through a periodic slab of meta atoms. After doing so I take the s parameters and calculate the linear parameters, Epsilon, Mu, n, z, etc etc.

This is fine for the linear case with the standard meta atom with no nonlinear inclusions, but I need to include a varactor diode (for now just the linear regime of the diode will do) because I want to see how much an inclusion will effect the s parameters and therefore effect the epsilon, mu, z, n etc retrieved. Once I verify that the inclusion of the diode will not drastically change the resonance frequency (we have specific frequency regime waveguides in the lab so I have to make sure it doesn't go outside that range so we can experiment with it) then I can request a pcb fabrication company fabricate my design.

The above I was hoping to achieve via lumped element capacitance + resistor (I know this won't take into account diode voltage bias but it might have to do for now!)

I am also hoping to use the linear fields and a circuit analysis a la Poutrina et al (see analysis of nonlinear metamaterials, Poutrina et al) to create a nonlinear polarisation term + additional voltage source (hopefully via coupling with electric circuit physics in comsol within a time domain study, although don't know how to do that either!) to generate second harmonics.

I can use the above to plot the distribution and amplitude of the different frequencies involved in the nonlinear process.

Sorry for rambling!

Any advice would be greatly appreciated.