The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Phase Noise and Jitter Measurements >> ISF/PPV for a comparator/driven circuit
https://designers-guide.org/forum/YaBB.pl?num=1477570396

Message started by deba on Oct 27^th, 2016, 5:13am

Title: ISF/PPV for a comparator/driven circuit
Post by deba on Oct 27^th, 2016, 5:13am

Hi All,

For oscillators PNOISE analysis using spectre, if one uses the argument "ppv=yes" it saves the PPV waveforms which are accessed through the results browser. This is of immense help when trying to understand the noise sources.

But for driven circuit like comparator, the option "ppv=yes" is not saving the PPV waveforms. Is it expected for driven circuits?

Thanks

Title: Re: ISF/PPV for a comparator/driven circuit
Post by sheldon on Oct 28^th, 2016, 5:48pm

Deba,

Why do you need the PPV? The noise source will directly provide
the noise contributions? If you perform PXF/PAC you should be able
to get the transfer functions to the output?

In this example, you can see the PNOISE summary for a dynamic comparator. The input pair is the largest source of noise and the reset
switches are the second largest source of noise. You can use the same
methodology to simulate the sampling aperture of the comparator. So
not sure what value the ISF/PPV would bring.

Sheldon

Title: Re: ISF/PPV for a comparator/driven circuit
Post by deba on Oct 31^st, 2016, 12:43am

Hi Sheldon,

Getting the sampling aperture, or other noise modulating function require PXF analysis and back calculating the ISF/PPV from them. Please see the below paper

J.Kim etal., "Impulse sensitivity analysis of Periodic Circuits", ICCAD 2008.

But if the PPV values are directly available as in oscillators pnoise analysis, then one need not calculate the PPVs back from PXF.

Title: Re: ISF/PPV for a comparator/driven circuit
Post by sheldon on Oct 31^st, 2016, 4:54pm

Deba,

What are you looking for? PXF should give you all the
information you need?

Art

Title: Re: ISF/PPV for a comparator/driven circuit
Post by deba on Nov 1^st, 2016, 6:49am

Hi Sheldon,

From PXF you need to carry out the summation mentioned in Eq. (14) of the paper to calculate the ISFs.

J.Kim etal., "Impulse sensitivity analysis of Periodic Circuits", ICCAD 2008.

There are other simulations methodologies published to calculate the ISFs indirectly. Below paper mentions a time domain technique

M. Jeeradit etal., "Characterizing sampling aperture of clocked comparators"

But if the spectre saves the "PPV" waveforms then one need not resort to methods mentioned above. For oscillators, "ppv=yes" option saves the "PPV" waveforms. But for comparators/other driven circuits it doesn't save.

My guess is for oscillators one is looking at the amount of final phase shift due to an impulse arriving at time τ. That is Γ(τ) = h(t0,τ) = h(∞,τ). But for comparators, t0 is of finite value.

Title: Re: ISF/PPV for a comparator/driven circuit
Post by sheldon on Nov 18^th, 2016, 10:13am

Deba,

The ISF is the time domain transfer function so PXF will give you
all the information you need. Next, Dr. Kim's method requires that
you configure the comparator in a "meta-stable" state while performing
periodic steady state analysis. The requirement for meta-stability is
inconsistent with periodic steady state. Meta-stability implies that the
circuit has not reached its steady-state.

Sheldon

Title: Re: ISF/PPV for a comparator/driven circuit
Post by deba on Nov 19^th, 2016, 8:03pm

Hi Sheldon,

Thanks for the help. PSS/PNOISE seems to be working for the comparator noise analysis. For my comparator, I can get the noise summary. Using that, I can optimize it. Below paper is using the similar method.

Noise-aware simulation-based sizing and optimization of clocked comparators, Rabuske, T. & Fernandes, J. Analog Integr Circ Sig Process (2014) 81: 723.