Stats

  • Locked Locked
  • Replies 3
  • Subscribers 125
  • Views 2321
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Parametric Analysis + IIP3 (pss/pac)

itos

Hi,

I want to obtain IIP3 for different values of a design variable (e.g. a resistor) using parametric analysis. I successfully did this with a mixer and qpss/qpac.

I just had the following line in the ADE L Outputs window and swept the variable using parametric analysis:

ipnVRI((v("/vop" ?result "qpac") - v("/vom" ?result "qpac")) '(1 -2) '(-1 0) ?rport 50 ?epoint -80)

Now I want to do the same thing for an LNA (and hence use pac/pss) and suddenly it does not work any longer! :-( This is the line for the ADE L outputs:

ipnVRI((v("/vop" ?result "pac") - v("/vom" ?result "pac")) '-2 '0 ?rport 50.0 ?epoint -80)

But clicking on "Plot outputs" after the parametric analysis does not open a plot. I just says "Status: Evaluating" for a moment and then prints a single number (maybe the last result of the parametric analysis). No plot window opens, no message in the Log.

Is this a bug?

EDIT:


This is just CRAZY! I found out the bug and it's not about pss vs qpss. 2 Issues:

  1. A sweep within pss across Pin must be enabled. As a dirty hack, one must enable sweep, set number of steps to one and add the desired extrapolation point to "Add specific points"
  2. In PAC I had for the single frequency VAR("fin")+1M. Cadence dislikes that. So one must write e.g. 2001M. Similarly, refrain from putting a design variable into ipnVRI, e.g. ?epoint VAR("Pin"). Cadence will run crazy too.

However there is a new issue now: ipnVRI changes results significantly when chaning epoint whereas ipnVRICurves does not (as expected). Example:

ipnVRICurves((v("/vop" ?result "pac") - v("/vom" ?result "pac")) '-2 '0 ?rport 50 ?epoint -80)
ipnVRI((v("/vop" ?result "pac") - v("/vom" ?result "pac")) '-2 '0 ?rport 50 ?epoint -80)

both give me 2.85643 dBm. However, changing to

ipnVRICurves((v("/vop" ?result "pac") - v("/vom" ?result "pac")) '-2 '0 ?rport 50 ?epoint -60)
ipnVRI((v("/vop" ?result "pac") - v("/vom" ?result "pac")) '-2 '0 ?rport 50 ?epoint -60)

the first one gives me 2.8568 whereas the second one gives me 6.5767 dBm! Any idea how to fix that?

  • Hi Itos

    What version of virtuoso and MMSIM are you using?   The following commands will give that information:

    spectre -W

    virtuoso -W

    Are you using Shooting Newton or Harmonic Balance?

    Are you getting any errors/warnings/messages in the CDS.log or spectre.out file?

    It's difficult what to say what the issue could be without more information.  You may want to contact Customer Support http://support.cadence.com for assistance.

     

    best regards,

    Tawna

     

     

     

  • Hi Tawna,

    Thank you! We just had a race condition ;-) (I edited my post with a solution while you were answering).

    $ spectre -W
    sub-version  13.1.0.186.isr5
    $ virtuoso -W
    sub-version  IC6.1.6-64b.500.11

    I tried both (hb and Newton), no messages in either (simulations run successfully). But as mentioned, there seems to be a bug which can be worked around by doing what I described.

    However, maybe you have an idea about the discrepancies between ipnVRI and ipnVRICurves?

  • Hi Itos, can you contact Cadence Customer Support with this question? I don't have time to set up a testcase and test this out right now. best regards, Tawna

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.