On Mon, 22 Dec 2003 14:59:31 -0800, Jim Smith wrote:
>Could you explain the mechanism (or point me to a source) whereby the
>sheld is rendered useless by the presence of a drain wire at HF?
I can do both. It's called "shield-current-induced noise," (SCIN). It
was first described by Neil Muncy (ex-W3WJE) in an AES paper in 1994,
based on research he did only at audio frequencies. I've done more
research on it in the last year or so, publishing my research in two
AES papers -- one in March, 2003, presented in Amsterdam, and one in
October, 2003, presented in NYC. Neil's 1994 paper and my March paper
are available from the AES (www.aes.org). The October paper will be
there in a few months (for some reason, it takes a while after the
convention to get it there). My March paper extended Neil Muncy's work
to 4 MHz, the NY paper took it up to 300 MHz. Both of my papers include
lots of data on real cables, and show the effect of various cable types
on RF detection in equipment. They also document my test setup, and
report of field testing of cables and equipment when exposing it to ham
transmitters on Field Day, and to broadcast transmitters.
The basic mechanism is inductive -- when current flows on the shield,
it induces voltage on each of the signal conductors that run inside the
shield. If the current is equally distributed over the shield, it will
couple relatively equally to the two signal conductors, and the
differential input will see zero volts (or nearly zero) due to that
current. But if the shield current couples more closely to one signal
conductor than the other, the two voltages will NOT be equal, so they
won't cancel.
In braid-shielded cables, especially those without a drain wire, the
current is quite evenly distributed over the shield. But if there is a
drain wire, the current will generally divide between the drain and the
rest of the shield based on the resistance of the drain and the total
shield. My research shows that below about 10 MHz, most of the current
in a foil/drain shield flows in the drain. As frequency increases, skin
effect causes the current to be more evenly distributed, and my
research suggests that braid-shielded cable loses its very advantage at
roughly 20 MHz (this will, I think, vary with cable construction).
Now, consider that in most foil/drain cables, the drain wire is twisted
at the same rate as the signal pair, and is usually constructed so that
the drain wire is much closer to one of the signal conductors than the
other. This is the principal cause of SCIN.
Braid-shielded cables do have SOME SCIN, but it is typically 30 dB
lower in level than foil/drain cables below about 2 MHz, about 20 dB
lower below 4 MHz. and about 10 dB lower below 8 MHz.
In the last ten years or so, cable manufacturers have been adding drain
wires to braid-shielded cables. My research shows that below about 2
MHz, the drain wire significantly degrades the SCIN performance,
approximately in proportion to the resistance ratios noted above. The
degradation is most severe below the AM broadcast band, and gradually
fades away as skin effect cause the current to flow more uniformly over
the total shield.
Without the effect of the drain wire (i.e., no drain wire or at higher
frequencies), the magnitude of SCIN is related to manufacturing
tolerances in the cable. That is, how equal are the lengths of the two
conductors, how symmetrically are they twisted, etc.
My research suggests that the ideal shield is a foil shield with a
light braid, much like is commonly used with MATV coax. The braid
distributes the current uniformly at all frequencies, and the foil
improves the percentage coverage. I've tested a rather low quality
foil/braid shielded cable in the 4 MHz project, and the results are
quite good. I didn't have time to include it in the 300 MHz work.
Unfortunately, there is virtually no balanced cable manufactured in the
US with foil/braid, or even braid only construction that is NEC rated
(safe for flame spread and noxious fumes) for permanent installation.
But Belden and Gepco are both aware of my work, and it is my
understanding that both are studying how they ought to respond.
BTW, Henry Ott has read my two Amsterdam papers which present the work
up to 4 MHz, and says that he agrees with my conclusions. I don't know
if he has seen the NY papers.
Jim Brown K9YC
http://audiosystemsgroup.com
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|