I have a board with 2 TPS54291 step down circuits to generate 4 different levels of supply lines, all in the range of 3.3V down to 1.2V. As input voltage, both circuits operate with 15V.
Both circuits (and all 4 channels) exhibit the exact same behaviour: the output voltage rises to the expected value set by the feedback network and it stays constant for a very short while (less than 1 mS) before it slowly decades to almost 0V. This cycle repeats about every 20mS.
What I have tried/checked:
* The duty cycle is always less than 40% and since the input to output voltage difference is so big, I assume it is not hiccup (a.k.a. UV) mode.
* Varying the load of the 4 lines, both increasing and decreasing it, the exact same wave at the output is seen.
* BP voltage is constant at 5.2V and the bootstrap capacitors also seem to have a fairly constant voltage (with the predicted spikes due to the higher mosfet conducting).
* By applying a series resistance to the inductor as a means to verify the possibility of overcurrent detection being the culprit, one can see that with 10 Ohm, the peaks are always lower than 0.5 A, thus never tripping the 2.2A or 3.8A limit.
* I also tried reducing and increasing the output filter cap: from 5uF up to 40uF. To no avail.
So what do you guys think I should try next? I've been on this issue for a while, so whatever suggestions or thoughts outside-the-box would be welcomed!
