I have simulated a simple LC VCO by Hspice, which oscillation frequency is 2.5G and I perform a fft simulation like:
.fft v(vco) start=100n stop=150n
but the probe-out spectrum is strange because the phase-noise is about -5dBc/Hz@1Mhz, this value is different from the result got out of SpectreRF of cadence(it is about -120dBc/Hz@1Mhz).