Connect with us

Multisim & PSPICE - using MOSFETs

Discussion in 'CAD' started by ER Yost, Jun 27, 2006.

Scroll to continue with content
  1. ER Yost

    ER Yost Guest


    I was wondering if anyone here has experience with modelling MOSFETs on
    PSPICE and Multisim. As a student, I've been used to using PSPICE 9.1
    Schematics to model and simulate circuits. However, I only have the
    evaluation version, and need more than 10 transistors in my circuit. I
    also have Multisim available to me, but PSPICE and Multisim seem to be
    very different in the way the MOSFET models operate. I took a rather
    uncomplicated N & PMOS from PSPICE and have gotten nowhere in Multisim.

    I'm wondering if anyone here has experience with trying to translate
    between the two programs, especially using transient analyses or the
    scopes that Multsim provides. Thank you.
  2. Jim Thompson

    Jim Thompson Guest

    What MOS model LEVEL=[a number goes here] is it?

    ...Jim Thompson
  3. ER Yost

    ER Yost Guest

    So far, in Multisim I've used level 1 virtual enhancement mode MOS as
    well as two power MOSFETs (P and N-channel) that come with the student
    version - one by Motorola, one by International Rectifier - I think one
    is a level 1, the other is a level 3. Power MOS isn't the right
    application for my research, but it at least it will give me an idea of
    what is going on for a few periods before giving me a timestep error
    (right now I have the relative tolerance at 0.01 and ITL4 at 100, which
    Multisim recommended for timestep errors).

    Another thing I'm trying to figure out in general is how to establish
    the channel resistance in the model since I'm looking into energy

  4. Hello ER Yost,

    have you ever heard of LTspice. It's practically free, unlimited
    and highly compatible to PSPICE on the netlist level. Your MOSFETs
    will work with LTspice if they work with PSPICE.

    There is also a user group.

    People should be aware that the GUIs of real instruments are mostly
    compromises because of limitations in the measurement hardware,
    limitations of the display, knobs and the cursor control.
    Why want you go with the second choice? I wonder.

    It's much more efficient and powerful to work with pure xy-diagrams
    in the waveform viewer of PSPICE(Probe), LTspice and many other SPICEs.

    Best regards,
  5. ER Yost

    ER Yost Guest

    Actually, yes, I downloaded LTspice yesterday and it has been very
    helpful. I've been able to build bigger circuits and get good feedback.


    With the research that I am doing this summer, it will be extremely
    important to be as accurate as possible. You can Google "quantum-dot
    cellular automata" (QCA) if you'd like - I'm doing research for Notre
    Dame (1st hit). The circuit that I am building isn't really to help the
    QCA project itself, but to provide a concrete example of how QCA is
    going to be more energy efficient than typical circuits today.
    I need something more accurate than LTSpice. Sure, PSPICE or Multisim
    aren't perfect, but they're pretty close. I'd love it if I could get a
    good enough version of PSPICE to have more than 10 transistors, because
    I have more experience with that. But, for now, I'd love to be able to
    use Multisim since it will give me enough computing power to do it.

    Nothing against LTSpice. It's definitely saved me time. It just won't
    be up to par in the end.


  6. Jim Thompson

    Jim Thompson Guest

    I don't think there's actually any issue with LTSpice accuracy as long
    as you don't use some of its speed-up gimmicks.

    But Level=1 and Level=3 models suck the big lemon.

    Your issues probably lie there as well as with your inexperience with
    Spice engines and how to set time-steps.

    (I've been using PSpice since it was on DOS ;-)

    ...Jim Thompson
  7. ER Yost

    ER Yost Guest

    Humbled I am.

    Until you mentioned it, I didn't really realize that I could set the
    MOS models in LTspice. Duh.

    Although I still don't get why my basic SPICE model wasn't consistent
    between programs, I might as well give LTspice my undivided attention
    and leave it at that.


  8. Hello ERY,

    LTspice has normally "data compression" enabled. It's like a lossy
    data compression of the raw-file. You should switch it off.

    ..options plotwinsize=0

    With this command, LTspice will beat PSPICE and others regarding
    accuracy by decades!
    Sorry that this compression setting isn't so obvious for new users.

    Best regards,
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.
Electronics Point Logo
Continue to site
Quote of the day