Power limiting via soft start

[foxtrot]

I'm working on a design based upon the LM76005 (https://www.ti.com/lit/ds/symlink/lm76005.pdf) buck converter. I need to limit inrush current (or, ideally, inrush power) to a fixed value, and I was considering using output tracking (SS/TRK pin) to accomplish this, in conjunction with either high-side current-sense amplifier and some discrete op-amps, or something like the LT2940 (https://www.analog.com/media/en/technical-documentation/data-sheets/2940f.pdf). I'm not an analog guy, so I'm looking for the easiest approach that I'm least likely to mess up.

Thoughts, napkin sketches, etc. are much appreciated 🙏

[PCB.Wiz]

How precise do you need the peak current to be ?
Once a part has a SS pin, that's usually enough to control the slew rate, and via that, plus the (known) total system load capacitance, the peak current is defined.
It is not usually necessary to apply more complex current envelope tracking ?

[foxtrot]

How precise do you need the peak current to be ?

Realistically, ±5% would be fine, I suspect. The only problem seems to be that I can't set the soft start duration long enough to charge the output capacitance – which is very high (2mF), in this design. I think I could use tracking, but I'm not entirely sure I understand it.

[Berni]

I had a similar application where i was using a switching regulator to charge a 3F or so supercap.

The way i did it there was using an analog solution that amplified the signal from a shunt resistor and used that to inject current into the FB pin when above a threshold. Since the shunt was on the output this resulted in the switching supply acting like a constant current source at a few amps to charge up the cap in a nice ramp. Once the set voltage starts to get reached then the regular resistor divider on the FB pin takes over, so the current in the output drops and the extra current limiting circuit stops messing with the FB pin.

The circuit was as simple as just a cheep high side current sense amplifier chip, creating a voltage proportional to current, then just put a diode (actually a BJT wired as a diode for more stable forward threshold) from the sense amplifier to the FB pin. So whenever the voltage out of the sense amplifier is above Vfb+0.6V it would pull up on the FB pin and convince the regulator to reduce output.

I had the output current limited because i wanted to charge my capacitor as fast as the switcher could (the limiting factor being the internal switch current), but i am guessing you want to limit the current going into the input of the switching regulator. In that case you can just move the shunt over to the input and it should work fine. You also want to leave some space for loop compensation capacitors in a few spots, since doing funky things to the FB pin can sometimes cause things to become unstable and oscillate. Those caps let you wrangle the regulation loop back into stability by slowing things down.