Strange phase-noise plot in PSS

Hi,

I am simulating a crystal oscillator (Pierce-Oscillator with Inverter), very similar to Tawnas example in

https://community.cadence.com/cadence_blogs_8/b/rf/posts/new-mmsim-12-1-harmonic-balance-features-auto-tstab-and-auto-harmonics

I applied the settings as described in 

https://community.cadence.com/cadence_blogs_8/b/rf/posts/guidelines-for-setting-pnoise-hbnoise-sidebands-to-get-accurate-results

and in the document from the $CDSHOME

Spectre Circuit Simulator and Accelerated Parallel Simulator RF Analysis in ADE Explorer Workshop.pdf

In the attached picture I compared 2 simulation methods:

In PSS/Pnoise simulation I get a strange  Phase Noise plot (RED curve).

In HB/HBnoise simulation I get a different result, but with much lower noisefloor (BLUE curve).

Image may be NSFW.
Clik here to view.

I am wondering what I am doing wrong in the PSS simulation (I am using 80 sidebands for Pnoiose) -why this strange step at 100Hz offset ?

Why is the noise floor 15dB lower for HB simulation ?

Any idea ?

Image may be NSFW.
Clik here to view.