Stats

  • Replies 3
  • Subscribers 129
  • Views 140
  • Members are here 0

Documentation on noise separation - source output

sierrawang

Hi,

I was wondering if there is any good documentation on ADE noise analysis - noise separation results? Specifically about the "Instance Output" vs "Source Output" option when plotting sim results. 

For example, in the Source Output option, there is an rgbi contribution (which I assume is the gate resistance thermal noise), that seems to correspond to the transistor's "rgate" op parameter based on my calculation.

On the other hand, in the Instance Output option, there is an rg contribution (also gate thermal noise?) that seems to correspond to the transistor's "rgbd" op paramter.

The two noises are off by several orders of magnitude. Which one is the more "accurate" gate thermal noise? And why are they different?

Thanks in advance!

Parents
  • Andrew Beckett

    I'm not sure it's "Instance Output" that you're looking at. The various outputs from Noise Separation are:

    Output Type Description
    Instance Output The total output noise from a contributing device. The noise is output referred (i.e. as it appears at the output of the circuit)
    Source Output The noise from a specific contributing noise source on a contributing device. Also output referred
    Instance Source The total noise from a contributing device, at the source (so excluding the transfer function to the output of the circuit)
    Primary Source The noise from a specific contributing noise source on a contributing device, as it appears at the noise source itself (i.e. excluding the transfer function)
    Src. Transfer Function The transfer function from a given noise source to the output of the circuit

    Perhaps you are comparing Source Output to Primary Source? The Instance Output wouldn't show individual noise contributors (so you wouldn't see "rg" there).

    The precise noise contributors depend upon the model used - so bsim4 has one set, bsimimg another, and so on. Unfortunately these are not particularly well documented... (for older models the names were fairly obvious, but they are not clearly documented in the model documentation (if at all)).

    If you need more detail or clarification on the specifics you're seeing, I suggest you contact customer support (use the case menu to submit a support case after logging in).

    Andrew

Reply
  • Andrew Beckett

    I'm not sure it's "Instance Output" that you're looking at. The various outputs from Noise Separation are:

    Output Type Description
    Instance Output The total output noise from a contributing device. The noise is output referred (i.e. as it appears at the output of the circuit)
    Source Output The noise from a specific contributing noise source on a contributing device. Also output referred
    Instance Source The total noise from a contributing device, at the source (so excluding the transfer function to the output of the circuit)
    Primary Source The noise from a specific contributing noise source on a contributing device, as it appears at the noise source itself (i.e. excluding the transfer function)
    Src. Transfer Function The transfer function from a given noise source to the output of the circuit

    Perhaps you are comparing Source Output to Primary Source? The Instance Output wouldn't show individual noise contributors (so you wouldn't see "rg" there).

    The precise noise contributors depend upon the model used - so bsim4 has one set, bsimimg another, and so on. Unfortunately these are not particularly well documented... (for older models the names were fairly obvious, but they are not clearly documented in the model documentation (if at all)).

    If you need more detail or clarification on the specifics you're seeing, I suggest you contact customer support (use the case menu to submit a support case after logging in).

    Andrew

Children
  • sierrawang
    in reply to Andrew Beckett

    Hi Andrew,

    Thanks for you prompt reply. I double checked with two PDKs that I have access to - both had an "rg" contribution in the Instance Output results. Although it's coming from the "resistor" type, not the "bsim4" type (which would only plot the transistor device itself).

    Another thing I noticed when comparing the noise separation sim between the two PDKs is that, the list of resistor-type Instance Output contributions are different: one PDK only has rg while the other also has rb, rd, etc. So that's probably why BSIM4 doesn't document that well, as it is process-dependent too and it might come down to how the foundry models/reports it. If that's the case, I doubt if Cadence customer support will be able to help much. But I will do some more digging and see. 

    Thanks,

    Sierra

  • Andrew Beckett
    in reply to sierrawang

    Sierra,

    I was talking about the noise sources within the device model itself; if the model is a subckt model and there are additional components around the MOS device on top of the internal sources within the bsim4 etc, then these additional contributors will (of course) play a part.

    Those within the transistor compact model itself (bsim4) are not well documented; those coming from additional components within subckt models can't be documented by us as they are foundry/process-specific. Anyway, the key point I was trying to make is the distinction between the different outputs from the noise separation mode of noise analysis - hopefully that helped a little at least!

    Andrew

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.