Flyback and Forward converter simulation

Discussion in 'General Electronics Chat' started by TheStig, Apr 30, 2007.

  1. TheStig

    Thread Starter New Member

    Apr 30, 2007
    4
    0
    Hi all,

    I want to simulate a Flyback and a Forward converter. I've been trying, but there is a convergence problem in PSpice. I need the simulation to be quite long (1 sec or more) so all the transitory effects are stinguished. I've simulated a Buck and Buck-Boost and had no problem at all. I'm also a little lost finding a model for the transformer. I'd like to take the magnetization inductance, the turns and if possible the resistance into account. And the forward to have 3 wings.
    Also, the voltage controlled switch I'm using doesn't seem a very elegant solution to me, is there any cleaner way to do it? Screenshots of the circuits I'm simulating:


    Flyback:
    http://img339.imageshack.us/img339/738/flybackbr7.jpg
    (ignore the error, I just touched something right before taking the screenshot)

    Forward:
    http://img61.imageshack.us/img61/7059/forwardar4.jpg
    (that only has 2 wings, but I just wanted to make the simulation work well)

    What I'm trying to simulate are these models:
    http://img240.imageshack.us/img240/7986/flybacktzu7.jpg (Flyback)
    http://img128.imageshack.us/img128/4403/forwardtfi5.jpg (Forward)

    If you know any site that provides similar examples or maybe one with converter schematics files it would be appreciated, I tried Googling but found nothing.

    Thanks in advance.
     
  2. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    I investigated modeling transformers in spice, recently. I have a spice model for a simple power transformer, and the AC Mains, at:

    http://www.fullnet.com/~tomg/gooteesp.htm

    which automatically calculates all of the needed parameters from the results of basic measurements. Instructions for taking the measurements are also included.

    That model's transformer probably won't quite suit your needs, as is. But there is a PDF-file reference included, that might also have techniques that are applicable to the transformers you want to model. And there are other papers about modeling transformers with spice, which you can definitely find with Google, with a search string such as "model OR modeling transformers spice". You should also do some similar searches at http://groups.google.com .

    There are also discussions and links to resources about modeling transformers, at places like http://www.diyaudio.com .

    Regarding schematics and simulations for various converters:

    First, I'd go to the Linear Technology Corp website, at http://www.linear.com . They have lots and lots of application notes (and other on line publications) with many, many schematics for different types of power converters.

    Linear also gives away LTspice (Pspice compatible), with models for all of their power supply IC products, opamps, and more. It used to come with a built-in switchmode PS "switch selector" and PSU auto-designer/simulator, that would take your specs and design a circuit with any of an appropriate set of their ICs. That part might need to be downloaded separately, now, IIRC.

    There is an "LT-Spice" discussion group, at http://www.yahoogroups.com , which has a library section with many, many example circuits and models, and a searchable archive of all message traffic.

    National.com (and other IC manufacturers) also have many application notes with lots of schematics, etc, for various types of power converters. National also has an on-line power supply design and simulation tool that is pretty slick, if they have an IC that helps do what you need. It even does basic thermal analysis and gives a thermal "image" of the board that's automatically designed during the process.

    http://www.maxim-ic.com also has lots of appnotes about converters.

    Some of the best stuff I've ever read about power supply design were the Unitrode appnotes and seminar notes. It appears that many or most of them are still available on the web, at http://www.smps.us/Unitrode.html , which also includes this link for Unitrode appnotes: http://www.unitrode.com/products/apps_prt.htm .

    Here is a collection of PSU design and related links that someone posted on Usenet (about a third of the way down the page):

    http://groups.google.com/group/rec....read/thread/64f48c64b5b1a680/4154242c9ed887c7

    There might be some good links, here: http://www.smps.com/ , and here: http://www.smpstech.com/websites.htm .

    This guy is "a bit eccentric", sometimes, to put it mildly. But he really knows his stuff. Some of it is now on his website, at http://www.genomerics.org/ . He posts on sci.electronics.design (searchable archive and message-posting accessible at http://groups.google.com ) as "Genome".

    Good luck.

    - Tom Gootee

    http://www.fullnet.com/~tomg/index.html

    -
     
  3. TheStig

    Thread Starter New Member

    Apr 30, 2007
    4
    0
    Now that was very helpful. Thanks a lot. I'm still an electronic engineering student and English is not my mother tongue, so I was a little lost googling. Thanks again!
     
  4. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    You are quite welcome. Someday, you can return the favor, to someone else.

    Your English is quite good, actually. Keep up the good work. It might be very useful, until we all have to learn Chinese. :) (unless you are lucky-enough to already know Chinese)

    If you don't mind my asking: Where are you studying? What year are you in? What courses are you taking right now? (Sorry if I seem overly curious. Once in a while I like to compare how they're doing it now, or in different parts of the world, to how they did it when I was in studying EE. It sometimes gives me hints about what one of my frame-of-reference's frame-of-reference is, or something like that.)

    - Tom Gootee
     
  5. TheStig

    Thread Starter New Member

    Apr 30, 2007
    4
    0
    :) Yeah I can understand and express myself decently in English, but when you go technical, you don't learn certain things. For example, I had no idea before that DC-DC converters were commonly refered to as SMPS. And googling without knowing that kind of things is sometimes hard.

    I'm studing Tenchnical Industrial Engineering (Industrial Electronics as the speciality) in Alcala University (www.uah.es), I'm in the last year (the 3rd one) with the final proyect in the making (related to vehicle occupant sensing), and plan to study the Superior Electronic Engineering or an official Advanced Electronics Systems Master, I'm undecided.

    I needed the simulations for the "Power Electronics Complements" subject. It turns out only 3 people chose that one, so we have to make an essay on a subject instead of an exam (I chose isolated DC-DC converters) and there are no classes. The simulations are made in PSpice but as no one taught us how to use it (the only remotely similar thing we learnt was Matlab 4.0 student edition...), and there are no classes, I had to learn it the hard way: reading the user's guide. The convergence problems without apparent reason didn't help at all, so that's why I looked for help!

    I get a lot of "wows" seeing how much a lot of people know about it, I hope I can reach that level some day! Oh and the finished Flyback model was something like this:

    http://img70.imageshack.us/img70/489/flybacksimma4.jpg
    http://img402.imageshack.us/img402/8845/flybacksimcr7.jpg

    Very very simple but serves the purpose. Chosing a low integration step time really helped with convergence.
     
  6. gootee

    Senior Member

    Apr 24, 2007
    447
    50
    Thanks for the reply.

    Your schematic and simulation results look pretty good.

    You might want to try adding an RC snubber network, probably in parallel with D1. You could start with something like 1K or 470R in series with .001 uF, and then adjust the capacitance for the least oscillation in I(D1).
     
  7. TheStig

    Thread Starter New Member

    Apr 30, 2007
    4
    0
  8. RmACK

    Active Member

    Nov 23, 2007
    54
    0
    You're doing better than me. What pspice are you using?
    I've had a lot of trouble with convergence errors in orcad 9. I can make a small change to cause the error then undo the change and still have an error after rerunning!

    Using a transformer was always a sure fire way to guarantee a convergence problem in my experience. I tend to model each side of the transformer separately -an inductors, resistors etc for primary and inductors, resistors & voltage sources for the secondary.

    The demo of orcad 8 never had convergence issues but the node limit & tiny component library are a problem. Sometimes I use this for little bits of a circuit but 9 has more realistic components in the library.

    Let me know how you get on with your simulations.
    I am trying out Simetrix soon as even trying to decipher what products Cadence/Orcad even have these days on their website seems a difficult task
     
Loading...