Post Go back to editing

LTC3805 issue in boost configuration

Hello everybody,

I'm working on a boost converter, used to boost a 400V input voltage up to 550V.

The circuit is build around the LTC3805 but I figure out mainly a startup problem: after the controller issues 4-5 pulses, it stop to work. After some seconds, it tries to restart, stops again and the sequence repeats indefinitely. Reading the datasheet, seems to be a hiccup mode triggered by an overcurrent; but this should not be as the OC pin is directly connected to GND.

What else may trigger an overcurrent or stopping the controller?

Power supply voltage and RUN signal are OK, the circuit is realized on a 4-layers board with all the GND pads directly connected to a ground plane.

The simulation of the circuit does not exibith any problem.

Any suggestion will be appreciated; thank you in advance,

Marco.

 tubes_boost.asc

  • I couldn't find the low pass filter of the ISenes pin 7 in DS or Evaluation boards. Essentially this circuit disables the slop compensation, DS page 10. Can you increase C14 to 0.1u and send to us the startup waveforms,  R89 and M7 drain?

  • Hello Victor,

    thank you for the reply.

    I added the 47R, 1nF network in order to reduce the noise on ISense pin (7), specially during the mosfet turn on. I also checked the circuit removing this network but I noticed no change. In any case there should not be need for slope compensation as the controller works always with D < 0.5.

    In the meantime I did some other tests and I found the circuit always starts moving the inductance a little away from the controller (I lifted the toroid about a centimeter from the pcb): doing so the circuit seems to work always in the right way.

    I suppose you mean C74 instead of C14 (it seems to be a typo error...): it is already 10uF.

    I'll be back with the waveforms as I return on this project.

    Thank you,

    Marco.

  • Hi Marco,

    You don't need too much cap at the soft-start SSFLT pin.  0.1uF is typical value.