Spice settings for Fourier Analysis

[SmileT]

I am goin to simulate an oscillator and i want to verify if the first armonic is at the right frequency...
If i want to have a resolution in frequency near 100hz for example is it right to set the end time of the transient simulation at 4*pi/(2*pi*100) considering the FFT as the convolution of a FT of a sine with the FT of a rectangular window?

I considered only the main lobe, so its amplitude is w=4pi/T for a rectangular window, and i wrote Min_res = w=4pi/T to obtain the above equation...
is it correct?

 

[SmileT]

Any ideas???

 

[MikeMl]

Spice's Fourier Analysis is useful for measuring harmonic distortion; it in not very useful for measuring the fundamental frequency of an oscillator. Why don't you do a .TRAN analysis long enough for the oscillator to reach steady-state, and just measure the period (time) of the oscillation? F=1/Period

 

[SmileT]

Spice's Fourier Analysis is useful for measuring harmonic distortion; it in not very useful for measuring the fundamental frequency of an oscillator. Why don't you do a .TRAN analysis long enough for the oscillator to reach steady-state, and just measure the period (time) of the oscillation? F=1/Period

Maybe i was not so clear... I am talking about the Fourier Tansform of the transient output; i simulate with .TRAN and i calculate the FT with the FFT option...
I am not sure of that: if i set the end time of transient simulation at T=4*pi/(2*pi*delta(f)) and do the FFT is it true that the resolution in frequency is delta(f) ?