By Bryan Geyer
A summary of the means used to interconnect audio equipment chassis
In the beginning (and
for ~ 35 years thereafter) the adopted “hi-fi” interface standard was the
ubiquitous "phono plug," a.k.a. the RCA plug (and jack). This 1940
relic utilized a relatively small (undefined) diameter coaxial cable wherein
the insulated center conductor is used for the hot side of the signal and the
common (sometimes called ground) side of that signal is carried by a shield
surrounding the center wire. The shield is intended to provide EMI noise
isolation, and is generally braided, sometimes spiral-wrapped. It’s composed of
copper, or tinned-copper, or, infrequently, aluminum foil. The hot center
conductor is routed to the RCA plug’s pin contact, and the shield is routed
directly to the shell. An outer jacket provides physical protection. Such cords
are readily available in assorted styles and lengths, with various exterior Ø
dimensions, dependent on shunt capacity rating. The latter ranges from ~
15pf/ft. (special “low capacity”) to ~ 35-45pf/ft. (normal).
At some time in the early 1980s, an upgraded version of this RCA-type coax was introduced (as best I can recall) by Hitachi. In this new version the insulated hot conductor remains connected as before, to the pin of both RCA plugs, while a separate common/ground conductor (insulated) is routed along with it, inside the coax core, that connects with the outside shell of both RCA plugs. This wiring completes the intended signal transfer, and eliminates the risk that the impedance might vary if the common was routed only through a braid. A shield overlay is then added to repel EMI. To avoid creating unintended chassis-to-chassis ground loops, this shield is mated to only one RCA shell. That mated end then serves as the designated signal input side, so such cable is arbitrarily defined as being unidirectional*. RCA cords of this type are now prolific; there are many suppliers, e.g. Audioquest.
Any benefit conveyed
with this upgraded style of unbalanced coaxial cable is directly related to
length (cumulative net impedance). If the required run is fairly short (≤ 1
meter), the likelihood of benefit is minimal. For runs > 1 meter, this
unidirectional style coax might possibly be superior to regular braided coax,
but not nearly so nice as the balanced AES/EBU option that’s described next.
AES/EBU: In the early 1980s another form of analog
interface evolved, as jointly developed by the Audio Engineering Society (AES)
and the European Broadcasting Union (EBU). This cable was initially conceived
for use in professional and commercial service, where there are often many more
cables, frequent disconnects and reconnects, longer runs, and where extreme EMI
exposure (strobed lighting, motorized generators) might prevail. This new
standard utilized “Cannon barrel plugs”, a connector that was created
(in 1950) by James Cannon. We now know these as “balanced” XLR plugs**.
Although handicapped by an oversized Ø (~3/4”) that reflects its ~ 70 year old
origin, these XLR connectors can provide useful advantages over ordinary “unbalanced”
RCA coax cable; e.g…
…the hot and
common/ground polarity terminal connections now engage simultaneously.
…the latched nature
of the plug-to-jack mating is more secure; less likely to inadvertently detach.
…dependent on the
related circuit design (see footnote**), an impedance-balanced XLR
connection will facilitate full common mode EMI noise rejection. This
latter feature can be a vital asset when heavy EMI is prevalent, and/or when
the interconnect runs get lengthy (> 1 meter).
More recently, XLR
plugs have been embraced by high-end audiophiles as general purpose interface
connectors. This has come about despite ample evidence that basic RCA coax can
provide fully sufficient noise immunity in any normal residential setting when
the cable runs are ≤ 1 meter***.
Concurrent with
creating AES/EBU, a new digital data transfer standard was developed to
address the new digital media. This protocol is known as AES3, and it encompasses
three distinct means of digital (only) data interface:
(a) Balanced,
using a single XLR terminal (at each cable end) for full 2-channel
stereo digital data transfer. One cable carries both of the encoded stereo
channels.
(b) Unbalanced,
using a single 75Ω coaxial cable, normally with RCA-type connectors, and
generally labeled “coaxial S/PDIF” (Sony/Philips Digital Interface). Such cable
is also offered with BNC terminals, provided the BNC connectors conform with 75Ω
(not 50Ω) coax. One cable carries both stereo channels. The standard coax most
commonly used is type RG59A/U, which has a 22 AWG compacted-copper center wire.
Type RG59B/U, with a 22 AWG solid copper-clad-steel center wire, can also be
used. The latter has slightly lower loss, but less flexibility. Both have Ø ≈
0.24 inch. These are professional quality cables, as used by commercial
recording studios. They are available on-line, in various lengths, directly
from U.S.-based supplier L-Com, at prices that are far less (often some
80% less) than the selling price from retail high-end audio sellers’ sites.
(c) Optical
digital data transfer, via an optical fiber (usually plastic, preferably glass)
cable, using Toshiba’s F05 style “Toslink” terminals. Also known as “optical
S/PDIF”.
While all three of
these digital data transfer means are practical and effective, most people use
option (b), the unbalanced 75Ω coax, using either RCA plugs or (less
frequently) BNC (75Ω) connectors.
TRS plugs: Three conductor TRS (tip-ring-sleeve) plugs
were derived from the old (circa 1930s) 1/4” diameter two conductor telephone
switchboard plug. The third wire contact was added via an isolated ring, with
the intent that a single plug could then serve for (unbalanced) analog stereo
(left-hot, right-hot, plus a common/ground) connector applications. Smaller
variants (3.5mm & 2.5mm Ø) soon followed. This basic 3 wire plug can also
serve for balanced analog connections, where its smaller (than XLR) size
often proves advantageous. As with the XLR plug, dual TRS connectors are
needed for balanced analog stereo service. The TRS configuration is not
quite so fully shielded as XLR, but it’s generally fine for home stereo; also
for most commercial service. Full-sized (1/4 inch) TRS plugs are the normal
commercial standard for guitar amplifier inputs/outputs; also for some dynamic
mics.
A note about the
photos: These plugs (a plug
= male, a jack = female, a socket denotes either gender, but
is chassis-mounted) are shown side-by-side to give you a sense of the relative
size (big!) of an XLR plug. That size becomes X2 when utilized for balanced analog
stereo (it requires separate L/R feeds). You can appreciate why a smaller size
connector might be of interest when EMI is benign.
BG (September 22, 2019)
*There’s persistent
chatter on audio forums (and in Audioquest’s sales pitch) supporting the
assertion that wire exhibits directional properties. No proof of this belief
exists (just aural hearsay), and prominent engineers don’t concur. This folly
stems from irrational groupthink. Wire is not directional; its
conductivity is fully bilateral.
**An XLR connector is
inherently “balanced” by virtue of its three wire symmetry. However, this “balanced”
state does not always extend to the associated active circuitry. Sometimes
these XLR connectors are simply wired in direct parallel with unbalanced RCA
inputs, so no phase cancellation (noise) benefit is then derived. It’s far
better to connect the hot signal (XLR pin 2) to the RCA hot input, and then
wire the cold input side (XLR pin 3) in series with a selected (adjustable)
resistance that’s precisely equal to the measured input impedance appearing at
pin 2. This creates an impedance balanced input that’s fully consistent
with effective common mode EMI noise rejection. Do appreciate that this
impedance balance can be achieved with virtually any equipment, provided it has
an internal means (generally a simple potentiometer) to set the desired pin
3-to-pin 2 impedance match. In very rare and costly instances, entirely
symmetrical balance can be automatically assured (no internal adjustment
required) by implementing a fully complementary circuit design, with
precisely matched differential stages throughout, from input-to-output. Such
equipment might, or might not, truly exist; the aspirations of audiophiles know
few limits. In any case, fully balanced common mode EMI noise rejection can be
accomplished by either means, and both would be equally effective.
Fascinating read! You've managed to present complex ideas in an accessible way, sparking a deeper interest in the topic. Your insights are invaluable, and I'm excited to see how this discussion evolves. Thanks for sharing your expertise—looking forward to your next enlightening post!
ReplyDeleteXLR Cable Assemblies
Enrgtech