Hello,
I'm finding the TPS53355 very difficult to troubleshoot, since as an integrated PWM+switch single-chip solution a lot of the signals you'd normally probe are no longer accessible, and TI hasn't added any sort of status pins to compensate for this.
Specifically, I've got a TPS53355 configured for 12V in, 1.8V out. When the load draws more than about 14-15 amps the TPS53355 does not seem to be able to sustain the load for more than a few seconds. After a few seconds at this level of current draw the LL node suddenly ceases switching for almost exactly 3ms, then resumes immediately (by which time the output voltage has fallen to ~800mV or so).
For simplicity, I've removed the overcurrent resistor (to rule out OCP as the reason for shutdown) and set the device for FCCM mode (to make the scope plots easier to understand).
Soft start is set to 5.6ms, so the fact that the device comes back after 3ms ought to rule out undervoltage protection since the UVP delay in this configuration is much much longer than 3ms (datasheet says 64.540ms). I probed VFB directly to rule out OVP as the cause; also the output voltage does not appear to be driven strongly downward (as it would be in an OVP situation where the TPS had decided to latch on the low-side MOSFET) although since I can't probe the low-side MOSFET gate this is just a guess. The datasheet is a bit vague on whether or not the 64.540ms delay applies to OVP as well; if it does then that is definitely not the case.
I've probed VREG and it does not appear to drop below 3.95V, so that ought to rule out UVLO being the cause of shutdown. Just to be sure I also tried wiring VREG to an additional outboard 5V LDO with very powerful drive; no change in behavior.
All that remains is thermal shutdown, but the behavior doesn't seem to be temperature dependent -- the amount of current that triggers the problem does not seem to change when I add copper heatsinks and extremely powerful fans in an air conditioned room, nor when I remove all heatsinks and operate the device in a very warm ambient temperature. Is there some unambiguous signal that can be used to know for absolutely certain if thermal shutdown has occurred or not?
Overall I'm glad TI has built in so many protections in this device, but I am very frustrated that they have not provided any indication of which one caused the shutdown. This sort of indication is pretty critical when troubleshooting a design.
So, the question: why has the TPS suddenly decided it is no longer in the mood to operate its high-side MOSFET?