What is the substrate definition procedure for a multilayer PCB in an EM simulator?

The substrate definition procedure for a multilayer PCB in an EM simulator creates an accurate representation of the PCB's layer stack that the solver uses to calculate the electromagnetic field distribution, impedance, loss, and coupling for all structures on the board. The procedure is: obtain the exact stackup from the PCB fabricator (request the detailed stackup drawing that specifies: each copper layer's thickness (typically 0.5 oz = 17 um, 1 oz = 35 um, 2 oz = 70 um), each dielectric layer's thickness (core and prepreg thicknesses, measured after lamination), the dielectric constant (Er or Dk) and loss tangent (Df or tan_delta) for each dielectric layer at the simulation frequency, and the copper surface roughness (Rz or RMS) for each copper layer), enter the layer definitions in the EM simulator (in Momentum: open the substrate editor and define each layer pair (conductor + dielectric) from bottom to top; in HFSS: create the layer stack in the modeler; in Sonnet: define the box layers; for each dielectric layer: enter the thickness, Er, and tan_delta values; for each conductor layer: enter the thickness and conductivity; define via connections between conductor layers with the correct drill diameter and pad size), handle the prepreg layers correctly (prepregs have different properties from cores because they are partially cured resin + glass cloth that flows and changes density during lamination; the fabricator should provide the post-lamination Er and thickness; if only pre-lamination data is available: reduce the thickness by 5-15% and adjust Er accordingly), and add surface finish effects (ENIG, HASL, or OSP finishes on the outer copper layers affect the loss at high frequencies; ENIG adds a nickel layer (1-5 um) with lower conductivity than copper, increasing loss by approximately 10-30% at frequencies above 10 GHz; model the nickel layer explicitly in the conductor stack).