>> That is, start with just power amp and speakers.Uh, er, you don't LEAVE the system this way for listening to music. You stick your ear by the speaker to see if, in the original poster's case, the offending "radio channel" can be heard. If it can, then the interference MUST be solved in this configuration BEFORE it can be reliably solved in a more complex one.
Now, how does Milen007 get music to play thru the speakers? Run the CDP directly into the power amp?
the FM signals are somewhat AM. So, the possibility of catching a FM station thru Milen007's system is quite real.Nope. FM interefence on baseband audio electronics is manifested mainly has hum and buzz, that changes in intensity and timbre as cables are moved around.
>> very few balanced input stages have much common-modeWhatever, dude. Actually, 3KHz is even quite optimistic, especially for active balanced inputs -- most I've measured deteriorate up from 60Hz by at least a single-order slope.
>> rejection above 3KHz or so.
Another terribly ignorant statement!!
For more information on the subject, I recommend Bill Whitlock's excellent papers (available at jensentransformers.com), which I agree with the vast majority of. I also highly respect the work of Douglas Self and recommend reading all of his stuff . . . even though on this particular subject I differ with his conclusions.