• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 1891 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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/LM5116-12EVAL: webench cannot create design similar to TI's LM5116-12EVAL !

Vlad Blanshey
Intellectual 770 points
Part Number: LM5116-12EVAL
Other Parts Discussed in Thread: LM5116, LM5116EVAL

Tool/software: WEBENCH® Design Tools

I tried to run webench for a DCDC switching converter design based on LM5116 . My requirements quite similar to TI's evaluation board LM5116-12EVAL : input 60V to 84V max , I need less power (3A output) and I want 2 layer board instead of 4 in eval board - cheaper production. I run webench and it created a design with the following MOSFETs :

Q1  VdsMax = 120.0 V IdsMax = 37.0 Amp  and Q2  Vdsmax = 200.0V IdsMax = 35.0Amps!

Parts chosen from Infenion and not surprisingly for these specs price was not known and total expense not calculated. Compared to TI's active LM5116-12 design this MOSFETs do not make sense, so

To trick webench I entered the same input -output requirements as in LM5116-12EVAL and webench was unable to come up with design recommending to choose another chip! In other words according to webench TI's design LM5116-12EVAL should not work :)

TI's LM5116-12 board features 80V max input and 12V  5Amps output - it uses simple Vishay MOSFETS both Q1 and Q2 the same type rated 80V max , 22mOhm and 12.5A . If this design works well with MOSFETs rated at the same input voltage (80V) then my webench design should do the same based on the same controller chip. Is webench wrong? If it is, then please let me know what MOSFETS should be there for my 84V input and less power design. I think that 100V max rated MOSFETS and about 12Amps should be sifficient.

Please advice.

  • 0
    TI__Mastermind 40040 points

    Hello Vlad,

    Can you provide me with clarification of your specification please?
    I am not clear from above what your output current is.

    so far I am assuming the following.

    Vin min 60V
    Vin max 84V
    Vout 12V?
    Iout <3A (please clarify?)

    Also, I am assuming you would like the lowest cost solution?

    WeBench is a general tool taking into account variations in overshoot due to layout design.  As such it leans on the conservative design side.   As such it selects FETs based on this fact.  You can reduce the Vinmax to trick WeBench for this situation?

    Also, there is a dial that you can turn to optimize the solution cost, may I suggest you make use of this?

    There is also a feature that allows you to individually change components, but again please bear in mind, the tool will be somewhat conservative and ultimately limited in the breadth of options.  I hope this helps?


    David.

  • 0 in reply to David Baba
    Intellectual 770 points

    Thank you David for reply,

    my output regulated 12V loaded up to 3A max (theoretically) , I expect in real life ~1A to ~1.5A ( just my greed, having more amps I can later decide to share amps with another module which currently uses 1A from it's own batteries:)

    Trying to adjust dial for cheaper BOM did not change mosfets - the same Q ratings -200V 35A. Seems "conservative" could be understatement in this case ... for 84VIN max using 200V rated MOSFET? If such huge voltage overhead were indeed needed to prevent overshoot I'd think the LM5116-based solution cannot be trusted. BTW Webench provides also optimal PCB layout of the components to prevent this... even traces are given...

    Webench is too complex tool to expect everything perfect there - my hope here to get real estimate on MOSFETs rating... is my estimate correct in using MOSFETs with 100V max should be good (matching limit for LM5116 and same for caps) and what about amps limit? If Eval board has 12.5A limit should that be enough in my case?

    Thanks again

    Vlad

  • 0 in reply to Vlad Blanshey
    Intellectual 770 points
    Also if you could share your view on PCB layout design generated in webench - can it be trusted? I checked layout and seems to follow the LM5116 specs recommendations. ANd it is 2 layer PCB, not 4 like eval board.
  • 0 in reply to Vlad Blanshey
    TI__Mastermind 40040 points

    Hello Vlad,

    I agree that solely from a 200V VDS rating is more than you need.

    Regarding your question about using a 100V MOSFET.  A 100V MOSFET should be OK, but of course, you will need to test and verify this to be sure. 

    Regarding your question about the amp rating (IDS) of the MOSFET.  I recommend that you use power loss to estimate the suitability of the power MOSFET.  For a SO8 Power pak I recommend keeping the max losses of each FET to less than 1W.  Again you will need to verify your selection over temp to ensure its reliable.

     

    I hope this helps?

     

    David.

  • 0 in reply to Vlad Blanshey
    TI__Mastermind 40040 points
    Hello Vlad,

    I trust the Webench layout recommendation. It ultimately follows guidelines for laying out a SMPS. Please refer to the datasheet layout recommendations for verification.

    Hope this helps?

    David.
  • 0 in reply to David Baba
    Intellectual 770 points

    David, how to estimate power loss in mosfet knowing only voltage range and internal Ron resistance?  Duty cycle is not known - it is changing from the change in the load. Also how to estimate mosfet's max current since it is rated peak current and averaged power loss does not suggest what this current suppose to be - their relationship is subject of a duty cycle too...

    Can you give example?

  • 0 in reply to David Baba
    Intellectual 770 points
    Thank you David - your input on webench generated layout is encouraging. Sure I will check in more details following the datasheet.
  • 0 in reply to Vlad Blanshey
    TI__Mastermind 40040 points
    Hello Vlad,

    Please see the basic calulation in the link below.

    www.ti.com/.../slva390.pdf


    I suggest you use this approach as a "first pass", to selecting the top and bottom MOSFETs.

    Suggest picking a few options that look good on paper and then testing performance on the bench.


    Hope this helps?
    David.
  • 0 in reply to Vlad Blanshey
    Intellectual 770 points
    If mathematical estimate is not possible, will you say that finding proper mosfets is subject of guessing-trial-and-error ? If too many specs variables involved trial-and-error could be the only way to find proper components... Is that how TI engineers arrive at their 80V 12.5A mosfets choice ?
  • 0 in reply to David Baba
    Intellectual 770 points
    David, when I pressed "reply" button previously (question referred to the possibility that math estimate is not even possible), then I saw your post with link to calculating efficiency that makes my previous question irrelevant. Calculations provided in that link, although having nothing to do with choosing the mosfets, nevertheless are very good to the estimate of the intended mosfets power dissipation as you suggested. That link reveals even more opportunities if both top and bottom mosfets chosen are the same with the same Rdson. I can use the LM5116Eval board data as a reference to find out all "other_losses" which will be the same in my case or lower if my load current is lower. Indeed very good calc approach to get initial idea on MOSFETs rating.
    Thank you very much.
    (I will probably need more advice related to LM5116 converter for choosing inductor but it should be another thread)
    Vlad