The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> spectre ac simulation problem
https://designers-guide.org/forum/YaBB.pl?num=1267782335

Message started by subgold on Mar 5^th, 2010, 1:45am

Title: spectre ac simulation problem
Post by subgold on Mar 5^th, 2010, 1:45am

hi all,

i have a problem of a simple ac simulation in spectre (cadence version: 5.10.41)

if i sweep the frequency (as default) in ac simulation, and do a parametric analysis by varying a variable (say Vdc), the simulator gives the correct results.

however, if i choose to sweep "Design Variable" on Vdc and check a certain frequency point in ac sim, the result is totally nonsense, except the first sweep point. for example, if i sweep Vdc from 0 to 1V, the results is only correct when Vdc is 0. if i sweep Vdc from 0.1V to 1V, then the result is correct when Vdc=0.1V. the same situation happens, regardless the sweep range, always the first point is the only correct one.

does somebody know the reason and a solution on this?

thank you very much.

Title: Re: spectre ac simulation problem
Post by Andrew Beckett on Mar 5^th, 2010, 1:57am

Can you post the input.scs (at least the analysis statements and options at the end)?

Also, which exact version of spectre are you using ("spectre -W" will tell you this, as will the output log from spectre)? If you're using spectre from IC5141, that's not advisable as it's very old - spectre has been part of a separate stream, MMSIM (latest version MMSIM72) for the last 5-6 years - it's still in IC5141 for legacy reasons (it's not in IC61X though).

Regards,

Andrew.

Title: Re: spectre ac simulation problem
Post by subgold on Mar 5^th, 2010, 2:28am

Andrew Beckett wrote on Mar 5^th, 2010, 1:57am:

here is the simulation info:

simulatorOptions options reltol=100e-6 vabstol=1e-6 iabstol=1e-12 \
temp=27.0 tnom=27 homotopy=all limit=delta scalem=1.0 scale=1.0 \
compatible=spice2 gmin=1e-18 rforce=1 maxnotes=5 maxwarns=5 digits=5 \
cols=80 pivrel=1e-3 sensfile="../psf/sens.output" checklimitdest=psf
dcOp dc write="spectre.dc" maxiters=150 maxsteps=10000 annotate=status
dcOpInfo info what=oppoint where=rawfile
ac ac freq=10 param=swing start=0 stop=1 annotate=status
designParamVals info what=parameters where=rawfile
primitives info what=primitives where=rawfile
subckts info what=subckts where=rawfile
saveOptions options save=allpub

it is IC 5.10.41.500.6. but "spectre -W" tells sub-version 7.1.1.071. what does this mean? Due to certain reasons, i am only allowed to use this version for the current project. so there is no solution for all the problems?

thanks again.

Title: Re: spectre ac simulation problem
Post by pancho_hideboo on Mar 6^th, 2010, 2:18am

subgold wrote on Mar 5^th, 2010, 2:28am:

ac ac freq=10 param=swing start=0 stop=1 annotate=status

Try to set "restart=yes" like following, although it might be helpless.

Quote:

ac ac freq=10 param=swing start=0 stop=1 annotate=status restart=yes

Title: Re: spectre ac simulation problem
Post by subgold on Mar 8^th, 2010, 10:04am

@andrew:

our eda guys give me some code to enable 64-bit operation with current spectre version. by typing:
spectre -debug3264 -V
i saw some message like:
App name: spectre
App path: /programs/cds/MMSIM62/tools.lnx86/spectre/bin
OS is 64-bit capable.
The user has selected 64-bit operation via the environment variables.
but it still doesn't work.
i am not very familar with these software issues, so maybe i am gonna find another way of simulating the circuit rather than spending more time on manipulating the simulator, but thank you anyway.

@pancho

thanks for your suggestions, but as you expected, it didn't work.

Title: Re: spectre ac simulation problem
Post by Andrew Beckett on Mar 8^th, 2010, 2:07pm

If the version number is 7.1.1.071 then it's MMSIM71 (Update 1), so fairly recent. You're using it with IC5141 (for the environment).

restart=yes was unlikely to help, because it should restart the DC anytime the DC is impacted by whatever you're sweeping.

Running 64 bit wouldn't help either. Rather oddly that was picking up MMSIM62, which is an earlier version. And all that command does is check that it's going to run in 64 bit mode.

I suspect something strange must be going on - it could well be a problem with the setup, but I think the best bet would be to contact Cadence Customer Support at http://support.cadence.com and then they (we) can investigate in more detail, hopefully with access to your data to take a proper look.

It's not a symptom I've seen, though.

Regards,

Andrew.

Title: Re: spectre ac simulation problem
Post by Geoffrey_Coram on Mar 10^th, 2010, 7:06am

Another question might be: how are you viewing your output? Could it be that the waveform viewer doesn't understand a single-frequency ac sweep?