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.

TPS23756: Internal hotswap MOSFET damaged? Primary 24VDC /Auxiliary source with PPD pin

Part Number: TPS23756

Dear Darwin

With respect to:

"For the inrush spec you refer to, the PD must finish charging the VDD-RTN capacitance before it can do normal current limit operation; otherwise, the PD will go into thermal cycling. The converter or the load should not be ON at initial startup so the PD can get out of the inrush current limit state. However, referencing the adapters to RTN will bypass this (as another consideration)."

Please, can you confirm this:

1) If I don't need the PoE classification and I want to use the TPS23756 as a simple Flyback converter with wide range of input (10.8VDC - 50VDC), then It's better OR'ed my 2 power supply (priority and backup) bettween VDD1 and RTN with:

-VDD and VSS unconnected( left float)

-DEN, CLS and PPD  left float?

If it's true:

2) The TPS shoud be work instantly at power the circuit in VDD1/RTN (don't care with inrush current, thermal cycling, etc) ?

By the way, My TPS was working fine with VDD1/RTN (backup supply) but today in the morning can't work and Vc = 2.2 V... Vout=0V

But, by the night, I test again without load, and by a moment, the Vout=12V for a minutes... and then, Vout= 0 V, Vc = 2.2V

If 1) and 2) , were true, Vc always be ~9Vdc?

  • Sorry, I correct this:

    1) If I don't need the PoE classification and I want to use the TPS23756 as a simple Flyback converter with wide range of input (10.8VDC - 50VDC), then It's better OR'ed my 2 power supply (priority and backup) bettween VDD1 and RTN with:

    -VDD and VSS unconnected( left float)

    -CLS and PPD left float?

    -DEN tie to VSS

    -APD tie to VDD1 through Resistor divider (Rapd1 Rapd2) as Figure 33 of datasheet OR is sufficient with DEN tied to VSS
    In both case, T2P output is enabled?
  • Hi Juan,

    1. That's correct, if not using PoE, you don't have to use the pins that relate to PoE. VDD and VSS can be left floating. CLS and PPD can be left floating. DEN can be tied to VSS and APD can be tied to RTN. The real application for APD and DEN was adapter priority; however, since PoE is non-existent in your application, these pins don't really have to be used; key thing is that the pin is terminated VSS and RTN respectively as left floating will cause issues.

    2. That's correct, inrush is completely bypassed and the adapter will charge the bulk cap right away with no thermal cycling of the PD. With VC at only 2.2V when you input the adapter, it makes me wonder if the IC is damaged (like the internal startup circuit is folding back or no longer regulating). With VC at 2.2V, switching cannot start. You may want to check if when you plug in the adapter there is no overshoots on VDD1 that could be damaging the IC.

    Regards,

    Darwin

  • Hi dear Darwin...

    First, from that you post the confirmation,, I was updating my circuit, and the attached .pdf  is the actual circuit. TPS23756 NO POE.pdf

    After a while hours ... 

    If the TPS could start... here is the attached picture with VC=OK and TP15(Bias)=OK

    but the converter could stop at any moment. In this moment, Vc enter in the mode showed in the attached picture 

    I think is the hiccup mode and start an  very audible noise randomly.

    The TPS randomly can restart and "work normally" until again enter Vc and hiccup mode?

    The question is? What's happening?

    For TPS23756, I think the wave correspond to Vc try to overcome the VCUV = 9V and decresing by VCUVH = 3.5V with Freq Hiccup mode = 13.3Hz

    It should be remember that I don't want to use POE circuitry and I want to use the TPS23756 as a flyback converter with wide input voltage 10.8 - 57Vdc.

    This is the reason why I base my design in www.ti.com/.../tidu190.pdf

     because i'ts use the flyback transformer MA5281-BL specified to work at 10.8 - 57Vdc

    Can you help me to find the any error or improve the start-up... I research and I find this, but no try yet... D6 and L3 

    Additionaly, I attach .zip wave of other Test Points (TP) in my circuit when work fine.

    TPS WORK OK.zip

  • Hi Juan,

    The inductor/diode solution should be used for active clamp forward converters. Since this is a flyback, you should continue to use the diode/cap design.

    To start, can you try increasing R27 to 25k? Or increase C8? Also, is C13 and C13_cer both installed? If so, I recommend not populating C13.

    Thanks!

    Regards,

    Darwin

  • Hi dear Darwin...

    Thank you for clarify the difference between active clamp and flyback.. OK.

    "To start, can you try increasing R27 to 25k Or increase C8? ?"

    Yes. I tried with R27 = 24.9K (original value) and added to C8 +22uF to the system(total cap 1uf+22uF+22uF)... but it doesn't improve the start-up 

    "Also, is C13 and C13_cer both installed? If so, I recommend not populating C13."

    Actual, C13 = 470pF,

    C13_cer is not installed.... Then, you recommend populated 330pf instead 470pf, or neither?

    BTW, the Vc showed in the  attached picture of my scope corresponded to hiccup mode? and the audible noise is an effect of this mode?

  • Hi Juan,

    My comments above was to rule out that the VC cap was drooping due to the initial energy when switching starts in addition to the added load of the totem pole gate drive and T2P current.

    For C13, either cap can be populated but not both.

    With that, we can say the converter is probably in hiccup mode because of some overload on the output or a short in your PCB (the schematic looks okay). You may want to double check the connection/impedances on your board to confirm that the components are connected correctly, no shows on the board, and the correct component values.
    Thanks!
    Regards,
    Darwin
  • Juan, Has this resolved your situation?

    We think it is resolved but would appreciate your agreement.

    Thank you for interest in PoE products from Texas Instruments,
  • Hi dear Darwin,,,
    After of analyze the output circuit... . I suspect of electrolytic capacitor C16 - 39uF/16V.
    I think that voltage rated was not sufficient and any moment the capacitor was damaged causing shorted randomly. I replaced with 47uf/50V and I have running without fall in hiccup mode from yesterday...
    I think to place a TVS of 12V in the output, before the capacitors...

    I hope this resolve my issue.
    Thank you very much
  • Hi Juan,

    Yes in the reference design it was 25V so in your design 50V should be okay too. Were you able to measure an over voltage on VC? In any case, yes I have seen designs with protection on VC in the past. Be sure the voltage breakdown you choose is higher than the normal operating voltage on the VC rail (so that it does not turn on in normal operation and only clamps when there is a true overvoltage).

    Thanks!