How to measure ISF in LC-VCO ?

[mazelk]

isf nmf

Acording to phase noise theory (Hajimiri et al.), how to measure ISF (Impulse Sensitivity Function) in a complementary LC-VCO. ??

How to do it in a SPICE-based simulators (e.g. ELDO, HSPICE) ??

 

[bozzo]

how to obtain isf from simulation

Hello,

I suppose you want to measure ISF to get L(f) ....

with Eldo, the .SST oscil and .SSTNOISE
simulations enable to get L(f) directly,
without ISF measurements.

Bozzo.

 

[qslazio]

L(f) surely can not offer design insight about how to lower the phase noise.
So we need ISF as well.
I think at present we can only get ISF by simulation which you can refer in hajimiri's papers.

 

[bozzo]

Eldo RF provides output tables which contain informations to know which device or which harmonic is responsible for the output noise.
Each device is listed, sorted by its contribution, and for each device we may know which harmonic is responsible for which amount of noise. Also the contribution per harmonic is given.

From that you can get transfer function between each noise source to phase output, and
optimize your oscillator.

Bozzo.

 

[hrkhari]

Hi Guys:

I'm also looking into the methods of measuring the ISF and NMF (Noise Modulating funtion) in order to obtain the ISF effective in a LC tuned VCO. Pls do suggest the procedure in executing it in HSPICE or Spectre-RF, if possible. According to Hajimiri, the simulations were performed utilzing HSPICE. Thanks in advance

RGds

 

[rfsystem]

If I remember right is the ISF response measured by injecting a small pulse and comparing the phase with an unpertubated system. That should be done over a number of steps over a oscillation cycle. So you have to write a script which repeat a spice simulation where the current/voltage injection source is shifted in time and a measure statement is executed.