jkeny wrote:If I understand it correctly, Nige is tying all the battery grounds together - that effectively bypasses any isolation on the input.
To correctly implement isolation the ground on the "dirty" side needs to be isolated from the ground on the "clean" side
spent a while testing this couple days ago
which was basically using the isolators as intended having the i2s unit fully isolated keeping both gnd and +3.3v separated both sides
versus merging both 3.3v and gnd of both sides of the isolators at the battery terminals
initially found fully isolated a half hair better (this difference is sq is probably not worth chasing if your in any way normal) however in the past i found isolation of i2s lines not much use when everything's cleanly powered (clean both sides) so i started to look for flaws in my front end 3.3v cell/rail
i found that after adding an additional cell and changing some wiring i could no longer find any negative effect having both sides of the isolators merged at the batteries
it may not be absolutely accurate to say it bypasses any isolation on the input as nothing has changed in the isolator chips the only way to actually do that is to remove the isolators and bridge across them
the isolators are still doing what they do and the power supply to both sides are going though a heavy duty filtration system in the battery cells
question remains if the isolators are removed would it be better or worse?
im really not sure
im considering investigating it