• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 64 subscribers
  • Views 2056 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

WEBENCH® Tools/TPS54418: How to export the average model for loop response calculation with TIna or Allegro connector?

Frank Schneider
Prodigy 60 points
Part Number: TPS54418
Other Parts Discussed in Thread: TINA-TI

Tool/software: WEBENCH® Design Tools

Hi,

unfortunately the WEBENCH does not provide the ability to model a Type III compensation for the TPS54418 nor does it provide the option to edit the schematic.

So my question is, how to export an average model of the TPS54418, which can be used with either Tina-TI or the WEBENCH Allegro connector? (Is the PSpice Lite sufficient for using the connector?)

My intend is to overcome the limitations WEBENCH has for this part, as I do need very fast transient response (large signal analysis), while also achieve very low ripple and noise, as well as stable over the commercial temperature range. While the transient analysis can be done with the exported model, the loop compensation cannot.

Currently the results derived from WEBENCH gives very inconsistent results.

  • 0
    TI__Guru**** 191445 points
    Frank,

    I would just go ahead and used the ready to download models from the tolls and software page here:
    www.ti.com/.../toolssoftware
    There are both TINA-TI and pspice models for time domain (transient) and frequency domain (average) analysis. I use these tools often.
  • 0 in reply to JohnTucker
    TI__Guru**** 191445 points

    Frank,

    It looks like Webench will only export the transient model.  I think you will need to use the download models as referenced in the above post.

  • 0 in reply to JohnTucker
    Prodigy 60 points

    Hi John,

    thanks for the kind reply.

    As suggested, I downloaded the Tina-TI average model of the TPS54418.

    But there a some questions regarding this model, you might be able to clarify.

    As the test bench description say, I changed the parameter for L and Cout, as shown below:

    Also the same for the inductor macro:

    But for the TPS54418 macro it does not work, it stays always as 1µH. (Even if I change the value and re-compile without error.)

    This said, my results with this setup, are around half of the bandwidth calculated using the approach of the TPS54718 data sheet.

    For sure I 'am missing something important (:looking puzzled:).

    Therefore, any help is more than welcome.

    Best regards

    Frank

  • 0 in reply to Frank Schneider
    TI__Guru**** 191445 points

    Frank,

    I'll take a look at this when I get back to the office tomorrow.  I am more familiar with the Pspice working environment, but I am sure you can change the inductor value.  I don't think you need t delve into the macro to do it.

  • 0 in reply to JohnTucker
    TI__Guru**** 191445 points

    Frank,

    I was able to change the value.  See the enclosed detailed instructions:

    TPS54418 TINA.docx

  • 0 in reply to JohnTucker
    Prodigy 60 points

    Hi John,

    thanks for checking out.

    So it seems it is sufficient to only change the inductor value of L1 and the parameter "L=value" inside the parameter option of the average model.

    Therefore, inside of the average model macro is only kind of default value or some normalization I do not need to change.

    Don't get distracted by using scientific nomenclature, I have some issue with comma vs. dot representation.

    On some elements Tina-TI likes to have the comma, on some the dot and sometimes neither of those are working ...

    In addition, I guess the thing I 'am missing is, how to derive the slope compensation part of the TPS54418. It looks that the total gain saturates around 55dB and I found some heavy formula to calculate slope compensation but I doubt not all necessary information is given into the TPS54418 datasheet.

    Further. it looks like the Spice model slvm279 (not found on the Web Site) is hard to get and I do not know, if a PSpice lite is sufficient for the use of it.

    Hence, I wonder, if it is possible to get the power stage transfer function out of the Tina-TI average model?

    Currently I 'am working with a small signal model (analytical and by simulation --> same results) but neglecting the slope compensation. This leads to wrong assumptions and results.

    So my question is, how to include the slope compensation in modeling the output stage for my design.

    This said, a heads up is very welcome. :)

    Best regards

    Frank

  • 0 in reply to Frank Schneider
    TI__Guru**** 191445 points
    Frank,

    The average model for TPS54418 is a simple APC switch type. It is an emulation of the entire circuit. It is not modeled as the actual internal circuit of the device. We may have an unencrypted version. Let me check.
  • 0 in reply to JohnTucker
    TI__Guru**** 191445 points
    Frank,

    We don't have an unencrypted model available. You can examine the power stage, compensated error amplifier and total loop gain and phase independently in TINA-TI.