there are always different phase and group velocities in dispersive structures, and most non-uniform boundary structures are dispersive.
draw an equation line, then extrude a cross-section along, or just apply a thickness. for faster simulation use polygonal line without curved surfaces.

do not use polygonal approximation in eigenmode, when defining a small part of the spiral. because the partial section start and end may not fit if not cuvelinear. and they should fit tightly for producing correct eigenmodes (dispersion curve). That is a kind of topological phase effect which will always make cross-sections not fitting if approximating the spiral with uniform polygons.

P.S. the "light line" is the TEM wave dispersion cone in vacuum. The velocity which you are calling a "light velocity" is a phase velocity. And since it is a simple ratio of phase number, yes, it is linked to the resonance. But strictly speaking, the phase number(or wavelength) defines the resonance. Beware that in nonreciprocal media, the sum of both-way phase numbers defines the resonance. HFSS driven modal is capable of solving nonreciprocal media (i.e.broken time reversal, where broken spatial reversal is your spiral twist)