[RFI] PC Speaker "Buzz"

[Jim Brown]

On Tue,6/30/2015 5:53 PM, Paul Christensen wrote:
> If RFI ingress is occurring on the input
> to a PC *power-amp* speaker system, then probably the most effective and
> least invasive way to mitigate the common mode current is to wrap several
> turns of the input cable through #31 mix ferrite cores.  Wrap as close as
> possible to the PC speaker's amplifier input.   It's the *turns count* 
> that
> adds series common mode impedance.  Have a look at K9YC's materials in 
> the
> link below.  For several mix types, he shows the number of turns 
> required to
> achieve a target choking impedance for a given frequency. 

It is  major mistake to think only of the input -- it is VERY common for 
RF to be coupled onto loudspeaker wiring and from there to the output 
circuit of the power amp driving the speakers, and from there, via the 
feedback network, to some part of the circuitry where it is detected.

ALSO -- virtually every piece of modern gear comes with Pin One 
Problems, which is a very common cause of RFI.

Also -- ALWAYS use TWISTED PAIR for loudspeaker wiring, NEVER zip cord, 
glorified or not. Twisted pair is VERY effective at reducing 
differential mode coupling.

Note also that when I talk about a loudspeaker, I'm talking about a 
passive element -- the loudspeaker itself. In the pro audio world, a 
loudspeaker enclosure with a built-in power amp would be called an 
"active" or "powered" loudspeaker. Most powered loudspeakers are 
notorious for RF susceptibility, even the most expensive pro units. I 
used to go through the aisles at AES shows with a Kenwood VHF/UHF talkie 
and exposing the bad actors.

The common means of coupling to these active boxes are 1) Pin One 
Problems on audio leads AND on mains power leads; 2) obvious circuit 
layout mistakes, like breaking a "ground" layer under a trace to add a 
trace on the ground layer that first layout guy forgot; and 3) the 
obvious ones, differential coupling onto signal leads and 4) inadequate 
shielding. Capacitors across inputs and outputs affect ONLY #3.

Single turns through almost any core are useless below VHF.

73, Jim K9YC