I have an LM25576 powering a GSM modem and the transient response fades the rail too much on the 2A pulses during GSM transmit (pic 13). The problem is at the switcher's output and not due to Ohmic loss downstream to where the modem is. Circuit is below.
I found that a 470uF electrolytic at C12 improved things a lot (pic below), but adding two undoes some of the improvement. 470uF at C2 helped a bit though (pic below). I've tinkered with C11 at 100nF which was disastrous; tried a 3R and 470pF snubber to the switch node, and tried R1 at 22k to slow the switching frequency down but neither did anything. Any experts who can advise?
Trying to decide if I want to stick with this part, or go to something different. Might like faster for smaller inductor. 12V in, 4.2V/3A out if you have any favorites. Also wondering if a linear regulator might be better. This only powers the GSM (4mA nominal other than the odd transmit) and an MCU drawing around 20mA. It's plugged in, not battery. Just needs to handle the 12V in.
Pretty urgent for time on this...
Initial
470uF at C2
470uF at C12
Thanks!
