On Mon, 22 Dec 2003 20:38:31 -0000, David Robbins K1TTT wrote:
>This is probably more of a tvi reflector question... is there one of
>those, or is this place good enough???
>
>I get into my new dvd 5.1 speaker system. I picked up some shielded
>speaker wire and ferrites. But what is the 'correct' way to implement
>them? Obviously at the speaker end of the wires there is no place to
>ground the shield so that will have to float. Or is there something
>else I should do with it?? At the receiver end I can tie all the
>shields together and to the ground there. What about the ferrites?
>Would you put one at the speaker end? Just with the 2 speaker wires
>through it or with the shield also?? And on the receiver end, put the
>ferrites on just the 2 speaker wires between where the shields are tied
>to ground and the receiver, or before tying the shields together??
Some general thoughts from a pro audio guy who is also an EMC guy.
1. In the pro audio world, we have learned (the hard way) the critical
importance of using TWISTED pair for loudspeaker wiring. I have never
needed to SHIELD loudspeaker wire, but I have often had to replace ZIP
cord wth twisted pair to get rid of RFI. So if the cable you bought is
not twisted, it is probably a bad idea.
2. If the shield on your cable is a foil with a drain wire, it is worse
than no shield at all for frequencies below about 20 MHz. That is
because any current flowing on the shield in this part of the spectrum
flows almost entirely on the drain wire, which usually has the same
twist ratio (lay, in cable-speak) as the twisted pair, and lies closer
to one conductor than the other. In this condition, the current flowing
on the shield will be converted to a differential mode voltage on the
signal pair! Above roughly 20 MHz, the current distributes more evenly
over the shield, and this becomes less of an issue.
3. If you do have a shield, it should tie to the shielding enclosure of
the electronics. A shield that is not grounded does nothing useful.
4, The principal function of a ferrite is to choke off the current.
Depending on the frequency of the interference, you may need a few
turns or a lot, and the ferrite material you use should be chosen with
the frequency of the interference in mind. In general, the lower the
frequency of the interference, the more inductance, and thus the more
turns, you will need.
5. The place for the chokes is ALWAYS at the point where they prevent
current flow from the antenna (speaker wire) into some electronics
that can detect it. Assuming that the loudspeakers don't have
amplifiers built into them, the ferrite chokes should go as close to
the power amp output (in the receiver) as possible. BTW, it is quite
common for RF to get into the output stages and back up to the input
via the feedback loop. If the loudspeakers DO have amps built into
them, you want chokes at the input to those amp terminals.
6. You may also need chokes on one or more of the inputs to the
receiver, but I would do only what is needed. You'll find other uses
for any chokes that may be left over. I would start with the twisted
speaker wire, then try a choke on the speaker wire if there is a
problem. If you still have problems, try the inputs that are picking up
the RF. When you find the one(s) that are problematic, try removing the
choke from the output and see if you're still clean.
Expect different results on different ham bands. With no suppression at
all, my home stereo is clean on everything 80 through 440 with the
exception of 6 meters, even with 100 watts on an indoor 80/40 dipole
running within 5 ft of the equipment. But it picks up 100 w on 6
meters from an antenna about 35 ft away!
There is excellent tutorial material in the pdf catalog on the
Fair-Rite website, in addition to lots of data sheets on various kinds
of ferrites that they manufacture. You can't buy direct from them, but
they do sell through various parts houses.
Jim Brown K9YC
http://audiosystemsgroup.com
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|