positive pnmx encountered for push-push oscillator in ads

I'm designning a 60GHz push-push VCO in ads. When set the noise ports as differential nodes (@30GHz) in HB noise controller, correct phase noise (pnmx) is given, around -70dBc@1MHz offset. But when set the noise port as single-ended (@60GHz), i got positive pnmx value... I've changed the osc port direction and location, but they didn't work... Is this related to HB simulation setup? Anyone got any idea? Thanks a lot.

The problem is that the noise is computed around the fundamental frequency (@30 GHz in your case). So, as there′s no signal at this frequency (ideal push-push), the SNR is positive and does not signify nothing. I have the same problem and would like to know if it′s possible to get the phase noise around the second harmonic with Agilent ADS.
Anybody knows ?

See the followings.
The Designer's Guide Community Forum - phase noise of ring oscillator using spectre
The Designer's Guide Community Forum - phase noise of ring oscillator using spectre
The Designer's Guide Community Forum - Phase noise analysis of autonomous ring oscillator

Yes, of course it's possible.
Use "Noise Controler" instead of "Noise Analysis" of HB Analysis controller.

True, thank you for your help Pancho