• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 38 subscribers
  • Views 1249 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.

LM25117 WebTHERM simulation output issue

FZ37
Intellectual 450 points

Other Parts Discussed in Thread: TPS40170, LM25117

Two thermal simulations are listed below. In (1) the temp of M1 is missing but in (2) all the temps are returned. Please help, trying to pick which one for the board. Thanks,

(1)

Design : 4561043/373 LM25117PMHX/NOPB

LM25117PMHX/NOPB 22.0V-26.0V to 12.00V @ 10.0A

(2)

Design : 4561043/372 LM25117PMHX/NOPB

LM25117PMHX/NOPB 22.0V-26.0V to 12.00V @ 10.0A

  • 0
    Intellectual 450 points
    A quick update:
    I ran the same design with a different inductor and the thermal simulation did finish with a 120C on M1 which is kind of too high.

    Design : 4561043/377 LM25117PMHX/NOPB
    LM25117PMHX/NOPB 22.0V-26.0V to 12.00V @ 10.0A
  • 0 in reply to FZ37
    TI__Intellectual 2360 points

    FZ37,

    It appears that you are trying different selections of FETs and other components to see which combination you prefer.  As you note, the specific FET selection, inductor selection, and switching frequency make a difference in the temperature of each FET, and the overall efficiency.  You can see a calculated estimate of FET and IC temperature in the Operating Values view. The Thermal simulation offers a more sophisticated estimate of the component temperature based on board construction & layout.

    I'm not sure why your example #1 did not display the temperature for M1.  We are looking into this further.  I do see that the Operating Value for M1 temperature shows that it is running hot (around 115C), and the thermal simulation may have found it to be even higher.

  • 0 in reply to Wanda Garrett - WEBENCH Design Center
    TI__Intellectual 2360 points
    FZ37,
    Looking into the design further, it appears the selected inductor is very large, and overlaps M1 in the thermal simulation. When that happens, M1 temperature will not be displayed. We are working on a fix for it. For now, note that you can compare estimated temperatures using the Operating Values.
  • 0 in reply to Wanda Garrett - WEBENCH Design Center
    TI__Intellectual 2360 points

    FZ37, 

    If you do have overlapping components, as in this design, you can move the components for the thermal simulation. Prepare to run a new Thermal Simulation, then click the big green button to "Edit Thermal Simulation Copper Model." This brings you into the editing environment. (Note that any edits are for thermal simulation purposes only, and do not change the schematic or netlist.) 

    In the editor, there is a set of controls on the left side of the view. In the set of click-boxes under "Show:" click the one for "Move Components." Now you can move the inductor out of the way temporarily and move the M1 FET up a bit, out from under the inductor. Put the inductor back where it was. (You may want to move M2 as well.) Then "Save & Close", and "Run Simulation." When the simulation completes, you will see the temperature for M1.

    For more information about editing in WebTHERM, please see www.ti.com/.../thermal-simulation.page. 

  • 0 in reply to Wanda Garrett - WEBENCH Design Center
    Intellectual 450 points
    Wanda,

    Thank you for your responses!
    I had two similar designs from Webench for the final decision.

    (1)
    Design : 4561043/381 TPS40170RGYR
    TPS40170RGYR 22.0V-26.0V to 12.00V @ 10.0A

    (2)
    Design : 4561043/380 LM25117PMHX/NOPB
    LM25117PMHX/NOPB 22.0V-26.0V to 12.00V @ 10.0A

    In the operating values, there are only ~1 degree C difference on M1, which is very reasonable since the power losses are very close between the two designs. But in the thermal simulations the temperature difference jumped to ~18 degree C. I can see the copper sizes are different. I am not sure if any other factors play a more significant role here. Thanks,
  • 0 in reply to FZ37
    TI__Intellectual 2360 points

    Hi FZ37,

    This is an interesting comparison.  The designs are very similar in terms of power dissipation, yet the M1 FET junction temperature appeared much higher in the LM25117 thermal simulation than in the TPS40170 thermal simulation.  Your clue about copper sizes directly connects with what's going on. Yes, in the LM25117 PCB example, the copper area under the M1 FET is smaller, especially on the "south" side of the package (as viewed in the simulator).  If you go into the editor mode, and expand that copper area under M1, then run a thermal simulation, you'll find its temperature will drop.

  • 0 in reply to Wanda Garrett - WEBENCH Design Center
    Intellectual 450 points
    Wanda,

    Thank you for looking into it. It makes more sense. only thing is the default layouts are a lot different between two similar parts and designs. Thanks,