number of noise/freq pairs in pss/pnoise simulation

Dear laoudi,

laoudi said:

Can I simply do that by adding noise/freq pairs in analogLib voltage source as shown in the figure or is it another way?

It seems to me a more accurate method is to include the actual circuit creating the supply voltage with the oscillator in a pss/pnoise simulation. There are a number of ways to estimate the impact of noise on the supply to an oscillator on the resulting phase noise. Three of the most common are:

1. Include the actual circuit generating the supply with the oscillator circuit in a pss/pnoise analysis (most accurate, but can be very time consuming)

2. Perform a small signal noise analysis on the circuit generating the supply voltage for the oscillator to determine its output noise versus frequency. The contribution of the resulting noise to the oscillator phase noise can be readily calculated to determine how significant its contribution is to the oscillator has noise (very quick and provides a very good estimate to determine if method 1 is required)

3. Perform a transient noise analysis

In the metholology you are proposing, there are a few limitations as detailed in the Cadence support article at URL:

support.cadence.com/.../ArticleAttachmentPortal

Do any of these suggestions help answer your concern?

Shawn

Dear laoudi,

laoudi said:

Can I simply do that by adding noise/freq pairs in analogLib voltage source as shown in the figure or is it another way?

It seems to me a more accurate method is to include the actual circuit creating the supply voltage with the oscillator in a pss/pnoise simulation. There are a number of ways to estimate the impact of noise on the supply to an oscillator on the resulting phase noise. Three of the most common are:

1. Include the actual circuit generating the supply with the oscillator circuit in a pss/pnoise analysis (most accurate, but can be very time consuming)

2. Perform a small signal noise analysis on the circuit generating the supply voltage for the oscillator to determine its output noise versus frequency. The contribution of the resulting noise to the oscillator phase noise can be readily calculated to determine how significant its contribution is to the oscillator has noise (very quick and provides a very good estimate to determine if method 1 is required)

3. Perform a transient noise analysis

In the metholology you are proposing, there are a few limitations as detailed in the Cadence support article at URL:

support.cadence.com/.../ArticleAttachmentPortal

Do any of these suggestions help answer your concern?

Shawn

Hi Shawn,

 Thanks for your reply.  I still have some questions. I have read the article in this link: what are  the limitations?It Is not clear to me from this article that noise/freq pairs are not taken into account in a pnoise sim. 

From your answers,

#1 I agree this method is straightforward. 

#2 This is what I actually did. I run a small signal analysis  on the curcuit generating the supply for osc to get the noise. Is there a theoretical formula to determine how this noise affects phase noise/jitter of osc? 

#3 Is there any manual to see how to simulate jitter from a transient noise sim? 

Kind Regards

Dear laoudi,

In response to your additional questions ...

> I still have some questions. I have read the article
> in this link: what are  the limitations?

Did you happen to catch the following statement "Spectre does linear interpolation between noise data points.  In transient noise analysis, all noise sources are band limited by noisefmax." at URL:

support.cadence.com/.../ArticleAttachmentPortal

> #1 I agree this method is straightforward.

It is also the most accurate.

> #2 ...Is there a theoretical formula to determine how this
> noise affects phase noise/jitter of osc?

Yes. The phase noise due to supply noise is directly related to the output phase noise by its Kvreg where Kvreg expresses how the oscillator output frequency varies with supply voltage. I have attached a Microsoft Excel workbook I created a number of years back that provides a plot of the contribution of an oscillator's supply noise to the oscillator output phase noise for your inspection. You may find it of use. You may enter up to two sets of data of voltage supply noise versus frequency data, some information on your oscillator frequency and its Kvreg and it will produce a plot of the contribution of each voltage noise characteristic on the osciillator phase noise.

> #3 Is there any manual to see how to simulate jitter from a transient noise sim?

There is a relatively new ViVA calculator function to estimate phase noise from a transient noise simulation. Some details are provided in the following Cadence support documents at URL:


support.cadence.com/.../ArticleAttachmentPortal

support.cadence.com/.../ArticleAttachmentPortal

and the RAK at URL:

support.cadence.com/.../ArticleAttachmentPortal

I hope this provides some answers to your questions.

Shawn

vco_pnoise_vreg_ss_template_sml_111615.xlsx