ISF Function Extraction in Cadence Virtuoso

Dear asrf,

asrf said:

s there any tutorial which explains the process of plotting the ISF function for a certain oscillator ?

In addition to performing an ISF manually which I have done (there are several references on methods used to perform this series of analyses), there is a means to save the ppv response from which the ISF can be obtained. Referring to the attached illustration of the pss analysis "Options..." "Output" tab, I have checked the "save osc ppv" radio button.

If needed, there are many papers speaking to the relationship between the ISF and ppv responses, but the two essentially show the same information (i.e., sensitivity of the VCO to the phase of a VCO at which a disturbance occurs).

Is this along the lines of what you were looking for?

Shawn

Thank you for the effort. So after enabling ppv I will able to plot it in results ?

I think what you mean is first I have to obtain ppv and then plot ISF using ppv.

After saving where can I access to ppv file ?

I think what you mean is first I have to obtain ppv and then plot ISF using ppv.

After saving where can I access to ppv file ?

Dear asrf,

asrf said:

So after enabling ppv I will able to plot it in results ?

The ppv file used to be saved to the "raw" directory in the netlist directory. From an old SpectreRF design manual:

The PPV file is saved into the raw directory after simulation. For example, after running an
autonomous circuit named vco.ckt with the Save osc PPV option set to yes, the PPV file
is placed in the vco.raw directory. The PPV file is named analysisID.td.ppv.pss.

I believe you can also specify the file into which it is saved through an additional argument in your pnoise setup. There is a pnoise argument :

which specifies where the PPV is written. Within the pnoise dialog GUI, there is a dialog box for additional parameters. I am guessing this is where you enter the location for the file after the keyword "ppvfile".

I have not tried this - but hopefully someone will correct me if I am misleading you on the latter!

I will also admit to the fact that there is some confusion about the existence of the ppv file. In some manuals, the documentation indicates it is no longer supported. I would try the option in a test VCO pss/pnoise analysis (one that does not take too long!) to verify this. I will also try this again and let you know if I encounter a problem.

Shawn

Thank you for your effort I appreciate it.

Honesty, I dont know how to find this raw directory. 

Dear asrf,

After a few experiments, I found that including a path to a file to store the ppv  in the pnoise additionalParams dialog box results in a spectre "FATAL ERROR". If I eliminate the text from the addtionalParams dailog box, the simulation starts with Cadence Version 18.1.0.556.isr13 64bit -- 10 Dec 2019.

asrf said:

Honesty, I dont know how to find this raw directory.

The raw directory is in the netlist directory. However, in my test simulation, there is a ppv file in the psf directory. The simulation is not complete, but is in the pnoise segment of the analysis.

Shawn

Dear asrf,

Well, some good news. The pnoise simulation completed. There is no text file in the raw directory. However, there is a binary ppv file in the psf directory. VivA recognizes the ppv file and one can plot the resulting ppv. Each waveform can be exported to a file if you wish. I have captured the Results browser showing the ppv binary file with a plot of the ppv for the P and N rails of  a VCO output waveform. I also exported the two ppv time domain waveforms to a table from which they may be saved as a text file if you wish.

I hope this helps.

Shawn

Thank you.

Is my setup correct ?

Dear asrf,

asrf said:

Is my setup correct ?

I am not sure if you intended to, but you included two copies of your noise GUI and no copies of your PSS GUI. You also did not include any of the option panels.

I am not sure what specifically your requirements are. Hence, between not knowing your requirements and not seeing your full setup, I cannot speak for all the settings you do show. A couple of comments follow.

1. I do note that your maximum sideband is only set to 5. Is that sufficient to capture the harmonics in your VCO output waveform and the number of harmonics that will be significant for your technology? It seems rather low.

2. Your plot show output signals "vop" and "von". However, your pnoise GUI panel show the positive node as "vop" and the negative node as "gnd!". Why is the negative node set to "gnd!" in lieu of "von"? It appears your signal of interest is a differential signal, but your pnoise analysis is set up for a single-ended waveform.

Shawn

My schematic

PSS setup

PSS options

pnoise

pnoise options

PPV results