Stats

  • Locked Locked
  • Replies 2
  • Subscribers 125
  • Views 13286
  • 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

Illogical outputs from cadence (IC5141) ADE.

RAO VINAY
 Hello all,

I installed cadence in a fresh drive which included base of IC5141 and hotfix of IC5141 in a separate folders. As a supporting package, I have installed updates of MMSIM 10.1 and IUS 8.2 (& many other like Assura, pacific etc). For checking the installation, simple inverter is drawn in schematic editor and voltage sources are taken from AnalogLIB.  After setting up the schematic, I tried to simulate it in ADE (using UMC90nm foundry) by choosing the spectre simulator. It couldn’t simulate and threw the following error (This error appeared only if I am running workstation from hotfix but not with the base).

“*ERROR*  Errors encountered during simulation. The simulator run log has not been generated. Possible cause could be an invalid command line option for the version of the simulator you are running. Choose setup-> Environment and verify that the command line options specified in the userCmdLineoption field are supported for the simulator. Alternatively, run the simulator standalone using the runsmulation file in the netlist directory to know the exact cause of the error.”

This is not a serious problem and found the solution through googling. Solution was to empty the maxnotes & maxwarns field which is found in simulator options (ADE->Simulation->option->analog->simulator options). After doing this, simulation ran successfully but got an illogical output. This is the serious illogical result which I am not able to resolve it. Details are given below.

First I tried with DC analysis (source voltage varied from 0 to 1v) and tried to plot DC response from both source voltage and as well as from the output of the inverter. I attached those results with this query.  DC response (first row in result) from THE SOURCE IS NOT A LINEAR CURVE AS EXPECTED rather response is of some kind of noise (though I didn’t select any kind of noise options). One thing to observe is that, voltage coming in Nano volts.  Output response is fine but it is sweeping from 1 to -1 as against the expected 1 to 0.

Secondly I tried with pulse voltage source from again analogLib but got into trouble again. Here again output response is fine but plotted source voltage is coming in terms of micro volts.

Can anyone explain, how can I resolve it? Is it because of simulator/mmsim/analogLib/maxnotes settings?

Your help would be highly appreciated.

Thank you.
Parents
  • Andrew Beckett

    The only way that I can see that you might have hit the first problem (using the IC5141 hotfix with the simulator run log not being generated) is if you have done the following:

    1. You have the SPECTRE_FEATURE_FILE UNIX environment variable set pointing to the MMSIM10.1 installation
    2. You have the <IC5141instDir>/tools/bin and <IC5141instDir>/tools/dfII/bin before the <MMSIMinstDir>/tools/bin in your UNIX path (e.g. $PATH env var if using bash/ksh/sh or $path shell var if using csh).

    Only if I do both can I get it to fail. This is because the IC5141 hotfix version looks for the feature file to detect whether the spectre version being used supports -maxwarns and -maxnotes on the simulator command line - the (very old) version of spectre that is shipped in the IC5141 installation itself did not support this. So what has happened is that you've overridden the feature file lookup mechanism to point to a specific feature file which tells it that it does support -maxwarns and -maxnotes, but yet are using a version of spectre (found first in your path) which actually does not - so spectre falls over in a heap.

    You should NOT specify SPECTRE_FEATURE_FILE yourself except in exceptional circumstances (if a Cadence AE tells you to in order to solve a specific, usually short term, problem, or for evaluating some special new feature). So please make sure that UNIX env var is not set. 

    In addition, you should ensure that the MMSIM path is before the IC5141 path otherwise it will pick up the old version of spectre shipped in the IC5141 installation, which you don't want.

    The second problem is nothing to do with maxwarns or maxnotes. it's almost certainly that you have the circuit connected up incorrectly and (maybe) supplies are not connected. Either that or you've got the scale of the devices wrong. Another possibility is that you don't have the UNIX environment variable $CDS_Netlisting_Mode set to "Analog" - and so the netlist is missing out some of the default values.  Hard to tell given that you didn't show the actual schematic or netlist.

    What I'd suggest is that you try typing in the CIW cdsGetNetlistMode() and see if it returns "Analog". If it doesn't, you'll need to do:

    setenv CDS_Netlisting_Mode Analog

    or

    export CDS_Netlisting_Mode=Analog

    depending on whether you're using csh (in the first case) or bash/ksh in the second case, and do this before you start icfb. You may need to do a Netlist->Recreate in ADE after that to ensure that it recreates the netlist from scratch.

    If that's not it, I suggest you post the input.scs here (not the models, because they're not yours to post) - but quite likely it will be possible to see what's wrong (especially if this is a relatively simple circuit - you say it's an inverter).

    I'm having to do a lot of guesswork here...

    Regards,

    Andrew.

Reply
  • Andrew Beckett

    The only way that I can see that you might have hit the first problem (using the IC5141 hotfix with the simulator run log not being generated) is if you have done the following:

    1. You have the SPECTRE_FEATURE_FILE UNIX environment variable set pointing to the MMSIM10.1 installation
    2. You have the <IC5141instDir>/tools/bin and <IC5141instDir>/tools/dfII/bin before the <MMSIMinstDir>/tools/bin in your UNIX path (e.g. $PATH env var if using bash/ksh/sh or $path shell var if using csh).

    Only if I do both can I get it to fail. This is because the IC5141 hotfix version looks for the feature file to detect whether the spectre version being used supports -maxwarns and -maxnotes on the simulator command line - the (very old) version of spectre that is shipped in the IC5141 installation itself did not support this. So what has happened is that you've overridden the feature file lookup mechanism to point to a specific feature file which tells it that it does support -maxwarns and -maxnotes, but yet are using a version of spectre (found first in your path) which actually does not - so spectre falls over in a heap.

    You should NOT specify SPECTRE_FEATURE_FILE yourself except in exceptional circumstances (if a Cadence AE tells you to in order to solve a specific, usually short term, problem, or for evaluating some special new feature). So please make sure that UNIX env var is not set. 

    In addition, you should ensure that the MMSIM path is before the IC5141 path otherwise it will pick up the old version of spectre shipped in the IC5141 installation, which you don't want.

    The second problem is nothing to do with maxwarns or maxnotes. it's almost certainly that you have the circuit connected up incorrectly and (maybe) supplies are not connected. Either that or you've got the scale of the devices wrong. Another possibility is that you don't have the UNIX environment variable $CDS_Netlisting_Mode set to "Analog" - and so the netlist is missing out some of the default values.  Hard to tell given that you didn't show the actual schematic or netlist.

    What I'd suggest is that you try typing in the CIW cdsGetNetlistMode() and see if it returns "Analog". If it doesn't, you'll need to do:

    setenv CDS_Netlisting_Mode Analog

    or

    export CDS_Netlisting_Mode=Analog

    depending on whether you're using csh (in the first case) or bash/ksh in the second case, and do this before you start icfb. You may need to do a Netlist->Recreate in ADE after that to ensure that it recreates the netlist from scratch.

    If that's not it, I suggest you post the input.scs here (not the models, because they're not yours to post) - but quite likely it will be possible to see what's wrong (especially if this is a relatively simple circuit - you say it's an inverter).

    I'm having to do a lot of guesswork here...

    Regards,

    Andrew.

Children
No Data

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.