As far as I know, S Balance denotes the dissipated power from a structure's ports, and it is equal to unity for a lossless structure. I considered the following simplified structure: A smooth cylindrical waveguide where its middle section has a larger radius compared to that of the input and output ports. Therefore, mode conversion will take place and the incident mode (the fundamental TE11 one) would eventually be converted to other supported modes if the frequency spectrum is the appropriate one. I have simulated in CST this structure and calculated the S Balance. In the mode, no losses (PEC BCs) were considered. Since it's lossless, I would expect that Sb~1. However, there is a deep minimum in the Sb versus the frequency figure. This result confuses me. So, what is the interpretation of Sb