Quote:
Originally Posted by
Avro_Arrow 
That is a switch mode power supply.
It's also likely not isolated. (It could be, but they don't bother to link you to a datasheet, so there's no way of telling short of buying it and testing it.)
Quote:
Originally Posted by
Avro_Arrow 
A 24 ohm, 1/2 watt resistor and a 500uF cap at 16 volts should be enough.
Resistors in series with power rails...yuck. If this helps, you've done it by taking a noisy power supply and increasing its ESR, which is also bad.
I also question whether a corner frequency of 13 Hz is low enough. This doesn't work like the CMoy's input filter, where we're trying to eliminate DC, which is an infinite number of decades below the corner frequency.
If this wall wart adds audible noise to the output of the amp, it'll probably be in the 10 kHz sort of range. Say it's quiet, -40 dBFS relative to the loudest music you can get through the system. That's equivalent to quiet office noise. If we simplify by saying the fc is 10 Hz or the noise is 13 kHz, that's 3 decades, buying us only 21 dB of attenuation. That's still -61 dBFS, which puts it only at the "library noise" sort of level, far higher than we'd really like.
If filtering is the answer, I'd say CLC makes more sense. But magnetics strong enough to get enough attenuation aren't cheap or compact.
For serious attenuation, you have to go active. I'd be happier to see a cheap ($6) unregulated wall wart followed by a naked LM7824 than either passive solution.
Edited by tangent - 11/11/10 at 10:18am