• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 3 answers
  • Subscribers 62 subscribers
  • Views 271 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.

LM25149-Q1: Convergence problems in PSpice for TI

Chris Brawn
Prodigy 120 points
Part Number: LM25149-Q1
Other Parts Discussed in Thread: CSD19502Q5B, CSD19505KTT, CSD18563Q5A, CSD18542KTT, CSD18534KCS, CSD17579Q5A, CSD17505Q5A, CSD16342Q5A, CSD15380F3

I've got a LM25149-Q1 design that simulates using the generic Pspice MOSFET models but when I change the MOSFETS models into our favourite device (DMTH8008LPSQ) PSice for TI refuses to converge even after experimenting with the simulation perameters. (e.g. RELTOL, VINTOL, ABSTOL and the like.)

I've attached the simulation files below.

Can you suggest anything I can do to get it to converge please?

Pspice LM25149-Q1 6-Feb-24.zip

  • 0
    TI__Mastermind 30420 points

    Hello Cheis,

    please try the following, hope this helps.

    modify the tolerances under PSPICE>Edit Simulation Profile.

    1. Select Analysis>max step
      1. Make smaller example 20ns

     

    1. Select Options
      1. select “analog simulation”>General
    1. There you will see a list of default tolerance settings
          1. Suggest increasing
            1. RELTOL, or VNTOL and ABSTOL
    1. Increase to 10x the default settings
    1. If all else fails there is the ability to select auto converge within the simulation settings.  Select under analog simulation>Auto Converge> check the boxes you wish to auto converge..
      1. This allows the user to define a tolerance range for all the simulation settings and will auto adjust within the limits  specified to try and get the model running.

     David.

  • 0 in reply to David Baba
    Prodigy 120 points

    Hi David,

     I have tried:-

    Maximum step size = 20ns (and larger/smaller values)

    RELTOL, VINTOL, ABSTOL at 10x their default values

    Selecting all AutoConverge options (including Restart)

    ...but I'm still getting these convergence errors:-

    I've also tried several different MOSFETS from the Texas Instruments/Power management/MOSFETS library (e.g. CSD19502Q5B) and they all produce similar convergence errors.

    Cheers

    Chris

     

  • 0 in reply to Chris Brawn
    TI__Mastermind 30420 points

    Hello Chris,

    I tried to run the model you provided and confirm that it runs into convergence issues.  There appears to be many modifications to the test bench that we provide.  

    Suggest the following, download the PSPICE model from the product folder and use the N-channel MOSFET breakout device and modify it to the FET you prefer to use.  Change one think at a time and run.  Suggest running the file with the said breakout devices before making any other changes.

    Hope this helps.

    David.

  • 0 in reply to David Baba
    Prodigy 120 points

    Hello David,

     I've replaced the DMTH8008LPSQ component with the "power_Mbreakn" and it does simulate without convergence errors. However I can't see how to modify the power_Mbreakn to make it look like DMTH8008LPSQ.  There appear to be properties for NRB, NRD, NRG and NRS but I have been unable to find any documentation about how to relate these to datasheet parameters such as Rdson, Qg and the like. Could you tell me how to do this or suggest some further reading?

    Cheers

    Chris

  • 0 in reply to Chris Brawn
    TI__Mastermind 30420 points

    Hello Craig,

    I think you should model performance with spice model, as for losses and efficiency, this is best calculated using the calculator.  changing the MBREAK parameters are not straight forward.  you will have to add Propeties to the model the FET you want to use.  

    Another idea is to find a FET similar to that you are using, there are many TI FETs in the library pick the closes match.  Hope this helps.

    David.

  • 0 in reply to David Baba
    Prodigy 120 points

    Hello David,

     As you say not straight forward, I think I may have difficulty finding some of those values.

    I have tried a selection of the TI MOSFETs (please see below) both in my design and the original snvmc96.zip (LM25149-Q1 reference design) and they all failed to converge.

    I will try the calculation spreadsheets for the losses.

    Would it be possible to contact the people who originally created the LM25149-Q1 model to see if they can explain why it has convergence problems with "real" MOSFET models?

    Best Regards

    Chris

    MOSFETs I have tried:-

    CSD19502Q5B
    CSD19505KTT
    CSD18563Q5A
    CSD18542KTT
    CSD18534KCS
    CSD17579Q5A
    CSD17505Q5A
    CSD16342Q5A
    CSD15380F3
  • 0 in reply to Chris Brawn
    TI__Mastermind 30420 points

    Hello Chris,

    We will take a look at this in due course, thanks for the feedback, please use the Calculator to get your through the design phase as i cannot give a time frame for completion at this stage.

    Thanks.

    David.