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.

TPS40055 switching waveforms

Other Parts Discussed in Thread: TPS40055

2211.6501V1R3.pdf

Hi,

My project is:

Vin =24Vdc stabilized, Vout=12 Vdc Iout from 500mA (st.by) to 10Amp full load

I attach the schema of the board i'm debugging.

I checked the switching waveform at the input  of LC filter (out of half bridge) and I found that in no load or light load

i get a pattern of  24Vdc for 2us and 0Vdc for 1.6us and in this situation the Vout is corect 12Vdc

When I apply the full load (this means tha a square waveform 220kHz is applied at the H brigde in the upper side of the schematic

the Vout drop at 9Volt and the pattern of voltage at the out is 24Vdc for 2.24 us after  0Vdc for 1.32us after 24Vdc for 400ns and 0Vdc for 3,24us

this pattern repets itself till the load is in after come back to the previous.

 

 

  • Massimo,

     

    The waveforms you describe are consistant with the pulse-by-pulse current limit issue I described before.  400ns is the minimum pulse width tha an over-current condition can detect while 2.24us is likely the maximum duty cycle at 24V input given the UVLO programming.

    While I appreciate the schematic and layout to help solve your issue, I would caution you in the future - the TI forum is a public access forum and anyone reviewing the forum will be able to view your schematic and PCB.

     

    That said, I believe I know the source of your issue.  The TPS40055 controller uses a comparator between the SW Pin (Pin 13) and the ILIM pin (Pin 16) to detect over-current.  In your layout R8 and R7 appear to connect to the input voltage traces fairly far from the drain voltage of the high-side FET (Q1).  Any voltage differences between the drain voltage of the high-side FET and the VIN voltage sensed by ILIM appears to the IC as voltage drop across the high-side FET.  Under load, the seperation between the Q1 drain and the VIN connection point for R7 is effectively increasing the current limit.

     

    A couple of things you can do:

    1) Improve layout to provide R7 with a kelvin connection to the drain of Q1

    2) Increase R7 + R8 to increase allowable voltage drop before current limit

    3) Increase UVLO by increasing R23, R27 and R28.  This will increase maximum duty cycle

     

    You may also want to review the 15nF snubber capacitor.  This capacitor is charging and discharging from 0V to 24V and back to 0V each switching cycle.  This dissipates 2x 1/2 CV^2 * Fsw = 15nF * 24V^2 * 300kHz = 2.6W  typically snubber capacitors in the 500 - 2000pF are most effective, depending on the parasitics around the switching node.  The target is for the capacitance to be no more than 3x the parasitic capacitance at the switching node.

  • Peter,

    well, thanks for Your suggestion today i'll do the improvement,

    yesterday i increased the Rilim and now the system is tripping at 10Amps.but i loose always regulation at 6 amp this means that

    at six amp the voltage drop from 13Vdc to 11volt.

    I think that the sw pin is the 12 otherwise my schematic is wrong.

    Is there any chance to disable the UVLO check, this because i use a regulated input voltage and i fit in the schematic a

    voltage controller to stop the work of the board is the 12 volt is not within the limit

    last question, what's the transient time when from no load condition I apply a full load?

    Thanks for the suggestion of beeing safe when publishing schematics anyway the board is a little piece of an high voltage 100kW

    power supply and i can make it public without problem, my problem is that without that 12 volt supply the whole generator can't work.

    best regards

    massimo