your so called PIN DIODE model may be innacurate.
in particular, it might not be simulating the I region thickness propperly, and the resulting transit time effects from an RF signal
and yes, a limiter needs a DC return path, AND an analysis program that can handle DC bias levels and rectification