Yesterday late in the night i ran a Spice Simulation of the Filters for Analog Voltage and it showed that they are not suited to block noise in the audible spectrum (10Hz-20kHz). One option to design them better is the usage of low-noise LDOs with a high input rejection to efficiently “filter” the audio supply-rails. I am thinking of a device like the LP3982.
Simulation was done using the freely-available LTSpice-Software under WINE. Fortunately the used ferrite-beads from Würth are included in its standard library.
Simulations shows that the filter is just blocking Signals well over 100kHz – all the audible noise will pass this filter rather unimpressed. Also tweaking the L is not really helpful.
Summing up all the issues on the PCB – I decided to run a redesign of the PCB now – before we start the production of the 10pcs. batch. I think this is the smarter way to do it. Because i don’t know what will happen after the first 10pcs. are produced – The thought is something like that: If we don’t do it know, we probably will never do the redesign …