• State
  • Locked Locked
  • Replies 9 replies
  • Subscribers 64 subscribers
  • Views 537 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.

Hot LM25119 problem

Ben Wai
Prodigy 150 points
Other Parts Discussed in Thread: LM25119, LM5119

I built a LM25119 with webench assisted values.  Input =18V, output 5V and 12V

Outputs are OK but there is heat problem with the controller that it becomes hot(untouchable) in about 10seconds even at no load.  I dare not to let it on and load it for more than 30sec fearing component damage.

Vcc measurements are different: Vcc1=6.7V   Vcc2=7V  Not sure if this any hint.  waveform capture at HB1 and HB2 is attached.

  HB1

HB2

Please help and let me know where should I look to.

  • 0
    Prodigy 150 points

    More waveform capture is done the HO1 and HO2 are stange while LO1 and LO2 are pulse chains.

    Ouput: 5.16V    Upper trae: OL1    Lower trae:  OH1

      Output=12.23V  A: OL2  B:OH2

  • 0 in reply to Ben Wai
    TI__Guru**** 191445 points

    Do you have the thermal pad well soldered?  how much copper area is attached?

  • 0 in reply to JohnTucker
    TI__Guru 58171 points

    Hi

    Check if any dv/dt induced turn on at LO pins.

    Check your Vds max of MOSFET's and try  R-C snubbers from SW nodes to the negative connection of sense resistors

    Regards,

  • 0 in reply to JohnTucker
    Prodigy 150 points

    I have the controller bottom pad soldered to PCB bottom copper across whole PCB about  100x100mm

  • 0 in reply to EricLee
    Prodigy 150 points

    Hi, Eric,

    Thanks for advice on snubbers addition.  The MOSFET is IPD036N04 of Vds=40V.   MOSFET are not hot at all. 

    I add snubbers of 1000pF in series with 5Ohm resistor, controller becomes less hot and I could see the no load power supply dropped from 100mA to 50mA.  Please advice where I could find info on snubber optimum values calculation.   Further I notice the sample circuit at LM25119 datasheet are two series 10Ohm resistors at CS and CSG pins to Rs ends.  Should I follow suit?

  • 0 in reply to Ben Wai
    Prodigy 150 points

    Typo mistake

    Exact current at no load was reduced from 120mA to 80mA.  Should it be even lower at no load?

    Controller is touchable but still hot!

  • 0 in reply to Ben Wai
    TI__Guru 58171 points

     Snubber design is popular. We can find a lot of articles on the web which explaining the snubber design. It is basically try&error because SW node oscillation depends on PC board layout.

    The rules which I am using are

    1. Rsnubber < 1/(4*Coss*Fosc)  :   A resistor drop of Rsnubber should be small enough to handle ringing current.  

    2. Csnubber > 1*Coss~5 *Coss  : Too small Csnubber is affected by temperature&MOSFET characteristics.

    3. 4xRsnubberxCsnubber > 1/ (5*Fosc) ~  1/ (3*Fosc)   : I typically damp 3 ~ 5 cycle of oscillation.

    4. Power loss of Rsnubber should be checked after selecting Rsnubber and Csnubber. Cds can be used rather than Coss. Fosc can be measured on the bench without a snubber.

     

  • 0 in reply to EricLee
    Prodigy 150 points

    After trying R-C snubber values combination, the 1000pF and 10R works the best.  But still consuming 70mA at no load which is too much over the stated Vin operating current at 7mA.  Theremust be a leaking at controller!  Further, how an obtain Webench circuit simulation the I can verify the waveforms at various ciruit nodes?

    Attached is the schematic of the circuit please help comment.

  • 0 in reply to Ben Wai
    TI__Guru 58171 points

    Hi

    LM5119 requires the 7mA Vin operating current. I guess the rest of current is consumed by switching loss including the snubber loss.

    Light load switching operation highly depends on the leakage current.

    Regards,