Propagation loss models should use the correct frequency

Moved from Bugzilla, see https://www.nsnam.org/bugzilla/show_bug.cgi?id=2122