No - you need to keep the path between the input capacitors and the SEPIC as short as possible.
In your circumstance I would put a lowRDSon N-channel MOSFET in the negative of the battery and then control the gate with a switch. Probably also have some sort of soft-start (slug the rise time of the gate with a capacitor) - you won't have to slug it much to reduce the in-rush current massively. Probably make sure your controller IC is setup to come on after your input MOSFET is hard on.
Your input MOSFET will cost alot less than a physical switch that can cope repeatedly with the in-rush current.
But won't this make the sepic converter always on? I mean if I put a mosfet at the negative side of the battery, how can it be powered with the microcontroller? in this case, maybe you meant powering the mosfet with a physical switch to its gate connecting it to the battery?
The LT3757A has a soft-start pin, going beyond the 10nF capacitor will hinder its working as I tested it in simulation. So if we go with this solution, we might put a 10nf at the gate of the mosfet to make it a bit slower. I guess this is better than using an analog switch.
I tried playing around with soft-start pin of the sepic ic but couldn't do much thus i kept it as is (10nF).
Let's say we went with this inrush current as is, is it gonna be a problem assuming physical or mosfet switch?
Now with 4S pack the voltage is 16v, thus I need to lower it a bit to feed it into the isolated DC-DC converter module. Previously I used a boost to make it from 6-8.4v to 12v but now it should be from say 12-16.8v to 12v using a buck converter assuming 3v is the lowest voltage each cell can reach. If we go to say 2.5-2.7v, pack voltage is gonna be 10-10.8v so it is less than 12v! in this case the buck will output about 9v or even less. I need to solve this the cheapest and simplest way possible.
![Banging Head |O](https://www.eevblog.com/forum/Smileys/default/bangheadonwall.gif)