Part Number:LM3150
24VDC in, 6.6VDC/12A out, right out of WEBENCH. Instead of using one large 100uf ceramic output cap, I used five 22uf ceramics (1206 size) in parallel. Regulator failed to start. I'd get one or two pulses on the high side gate, then the high side gate stuck high for about 7us, then went low and stayed that way forever. Output voltage had just barely started to ramp up, then fell to zero.
I tried different add-ons of input and output caps to no avail. Finally took off the five 22uf ceramics and replaced them with a single 100uf electrolytic...and it started working. Pretty noisy though, so I added back one 22uf ceramic and that still worked. When I tried a 470uf electrolytic, it stopped working again.
I found no explanation for this in the datasheet; how is it possible that the oscillator stops working after a few cycles and never recovers? It appears to be something related to the output capacitor ESR, and I'm assuming that some part of the feedback loop is relying on more of a switching 'glitch' than was being allowed by the parallel ceramic caps. Strange that it would just halt and never come on. Assuming it was maybe something related to input power ramp-up rate, I put in a one second delay from the microcontroller and added a FET to pull down the enable signal until power-up was complete and stable. That didn't help. Screen shot below illustrates the behavior...24V is up and stable, and we're triggering at the end of the one second delay as enable is released high.
Yellow trace is Vin (24VDC) and blue trace is the high-side FET gate.
