EMI considerations on LC filter SMPS

[MacIntoshCZ]

I have two stage LC filter, by two stage i means this: https://ridleyengineering.com/images/phocadownload/1%20second%20stage%20filter%20design.pdf
Coil in first stage is placed on noisy node (100Khz 250V peak), first idea i got was to place inductors at "grounds nodes" (see picture to get idea what i mean by that, and ignore thats seconds stage inductor  ) but by doing so, feedback loop will mess up.
So this is probably not the way i guess. Or is that possible? I never saw that to be honest =D. Even most of application notes has coils next to cap+ nodes....
So probably the only way is to use shielded inductor. Dont know if i can rely on Ycap connected to the enclosure, since EMI is my invisible enemy number one =D.

[T3sl4co1l]

If ground is declared after the inductor, that's fine.  This likely precludes sync rect and common-ground (non-isolated buck/boost) applications.  But other than that, the loop doesn't know anything different.

I suppose I can assume isolated, since you mention Y cap.  Notice there is a common mode path through the series capacitors, so the Y cap can still be placed at the output ground side without much difference.

Back in the day, ground-side chokes were a common trick to alleviate isolation ratings in transmitters -- otherwise it would have to stand off the full, whatever, 3000V the transmitter is powered by -- way back in the tube amplifier days that is, where a 5kW transmitter might use 3kV 2A or so, into a forced-air or even water-cooled ceramic triode or tetrode.

Tim

[MacIntoshCZ]

You are right, thanks for noticing!
Yes, its isolated!

[MacIntoshCZ]

Do you think its necessary to have inductor isolated? Since it will be in conductive enclosure connected wit Y cap to primary and secondary ground, it should limit EMI to go outside, but inside enclosure it could affect near electronics.
I am trying not to place anything sensitive nearby. In worst case its peaking at 250V 100Khz. Its rise time/ fall time what matter the most