Cadence tool suggestion

Discussion in 'Cadence' started by Kuan Zhou, Apr 27, 2005.

  1. Kuan Zhou

    Kuan Zhou Guest

    Hi,

    I am currently designing a chip working under a noisy environment. For
    example, the chip is working in a place with lots of strong
    electromagnetic fields. These fields will impact my circuit
    performance definitely. Is there a tool in Cadence to analyze such effects?

    What are the tools to simulate and measure the temperatures and
    rise/fall times over the entire chip. I often saw pictures shown by people
    with blue regions for low tempearature and red regions for high
    temperature. But I never know the tools they used to do such things.

    Thank you in advance!

    Kuan
     
    Kuan Zhou, Apr 27, 2005
    #1
  2. Kuan Zhou

    G Vandevalk Guest

    I too would like to hear this.

    I have written a spectre/spice netlister that dumps out the location of each
    extracted component into x= and y= parameters. (used the locations that I
    placed there
    in the diva extract run !!!)
    Then each of the components were given an initial dtemp=0 parameter.
    (i.e. I had the processing group add dtemp to the temp parameter. )
    Then I run a "typical" simulation run. Then I process the power dissipated
    in each of the
    devices ( mostly npn's and resistors ) and I calculate how much each device
    would raise the
    local temp. (more of a guess based on location. ... ) and how much it would
    raise the temp of
    local neighbours ... ( based on proximity ... tailing off rapidly ... )
    Now I insert the calculated set of dtemp parameters back into the netlist
    and rerun the simulation.
    This results in a new set of dtemp parameters ... rerun this loop until
    convergence or thermal runaway ( some dtemp is >tmax ) or after several
    runs. (we may have some positive feedback in our crude dtemp model so go
    back & damp the dtemp change rate .... )
    This was plugged into a framework that would then display dtemps on devices.
    I have seen this turned into
    a thermal map ... YMMV

    Note that large magnetic noise is an animal of a different stripe
    completely.
    Typically the smaller the x-section -- the less the pickup .. Good luck in
    any modeling here.

    -- G
     
    G Vandevalk, Apr 29, 2005
    #2
  3. I don't know of a Cadence tool that does this in the CIC space (I believe the
    PCB tools have something like this).

    Perhaps Kuan was thinking of tools like VoltageStorm/ElectronStorm/SignalStorm
    which use a colour map to indicate areas of (say) electromigration problems.
    Or SubstrateStorm (now Substrate Noise Analyst) which does an AC analysis of a
    noise perturbation to show how it propagates to an area on the layout via the
    substrate.

    Regards,

    Andrew.
     
    Andrew Beckett, May 3, 2005
    #3
Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.