Stats

  • Locked Locked
  • Replies 5
  • Subscribers 126
  • Views 11476
  • 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

Spectre stability analysis with active device as probe

StephanWeber

Hi,

this is a very nice feature, because sometimes you cannot insert a probe into the schematic. And in the past it worked fine for me in most pdks.

However, now I have problems, that the active device is not recognized correctly. If I use it from GUI and cross-probing, I got this netlisted:

stb4 stb  start=1 stop=1G dec=10 probe=DUT.M41 annotate=status

But this does not work.

I believe the problem is that M41 transistor model is a subckt with device m1 inside which is the active element. So I changed the netlist to this:

stb  stb  start=1 stop=1G dec=10 probe=DUT.M41.m1 annotate=status

But I get the same error message in Spectre (remark: the M41.m1 syntax works fine for DCOp like to get vdsat).

Also this is not working:


stb4  stb  start=1 stop=1G dec=10 probe=DUT.M41.m1.gm annotate=status

So I run out of ideas, can anybody help?

Here are the error messages:


Error found by spectre during STB analysis `stb'.
    ERROR (SPECTRE-16418): Invalid probe instance. Possible choices are iprobe, vsource, or supported active device.

Analysis `stb' was terminated prematurely due to an error.

Error found by spectre.
    ERROR (SPECTRE-16850): Invalid instance name `DUT.M41.m1.gm' is specified as value for the 'probe' parameter in stb analysis. Ensure that the instance name is specified in the netlist or specify the correct instance name in stb analysis.

Analysis `stb' was terminated prematurely due to an error.

Error found by spectre.
    ERROR (SPECTRE-16850): Invalid instance name `DUT.M41' is specified as value for the 'probe' parameter in stb analysis. Ensure that the instance name is specified in the netlist or specify the correct instance name in stb analysis.

Bye Stephan

  • Are you sure that this is really what you are wanting to do? Pointing at an active device will use the loop within the device (which may it is what you want, I'm not sure).

    There have been some improvements via IC6.1.8/ICADVM20.1 ISR25 which make it easier to set up a stb analysis in ADE using the terminal of an instance (rather than just pointing at an iprobe/zero-volt source), or for differential/common-mode stability you can point at a pair of terminals rather than having to insert a diffstbprobe. 

    However, these do require that the instance and terminal are a leaf component because spectre needs to have a single device to insert the probe (in effect) in series with. Without knowing the structure of your devices in the PDK you're using, it's hard to give a more precise answer. I'd be quite surprised though if you are really intending to use the device loop approach though.

    Perhaps you should talk to customer support?

    Andrew

  • in reply to Andrew Beckett

    Hi, after some experimenting with using the results browser, I could guess how the stb statement should look like for my own verilogapmos  model.

    This was the solution:

    stb  stb  start=1 stop=1G dec=10 probe=PM0.I0.vpmos annotate=status

    However, with the 5V pmos in the pdk I still found no solution, and e.g. this still fails:

    stb  stb  start=1 stop=1G dec=10 probe=M0.m1 annotate=status

    I wonder, could it be that pspnqs103 model is not supported for device-based stb analysis?

    Bye Stephan

  • in reply to StephanWeber

    Hi Stephan,

    Without trying it, I don't know whether the device-based approach (i.e. for feedback loops within the device) is supported for pspnqs103 (I can't see any reports of issues with this though). If it's not working, please contact customer support.

    Andrew

  • in reply to Andrew Beckett

    OK. There are some advantages with device-based approach, e.g. for postlayout sims and if e.g. Cgd of a transistor is a main part of the feedback loop.

  • The STB analysis linearizes the circuit about the DC operating point and computes the loop gain and gain and phase margins (if the sweep variable is frequency) ...

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.