As has been pointed out already, star quad has higher RF rejection,
but at the expense of higher cable capacitance. That means decreased
HF response*, especially for longer cables. [*All cables with two or
more conductors create low pass filters when connected between audio
output and input circuits: the longer the cable, the lower the cut-
off frequency. Whether the cut-off frequency is in the audio
bandwidth is another question, of course. It is usually only a
problem for very long cable runs, exceeding 100m or so. As a matter
of interest, some high quality music recording engineers are running
their analogue signals down 110 ohm AES cable, due to its lower
capacitance for longer cable runs...]
Also, due to having four inner conductors rather than two, star quad
is more expensive, heavier, harder to roll, harder to lay, and harder
to assemble - I have seen well-intentioned novices make a right
proper mess of trying to solder star quad's two inner conductors into
one XLR pin. Combined with the difficulty of trying to jam four inner
conductors into the XLR shell along with the shield, the result is a
cable that is far less than reliable.
In news groups dedicated to high quality music recording and
mastering, many engineers have moved away from star quad designs for
all of the reasons given above. Unless you're working around radio
transmitters or similar, its extra RF rejection is simply
unwarranted - the disadvantages outweigh the advantages. However,
star quad cable is often used in live sound reinforcement systems due
to its more rugged construction and enhanced RF rejection. In those
situations, where the signal passes through considerable processing
and/or equalisation before being heard, a bit of HF roll-off is not
going to be a problem.
A good balanced connection relies on two forms of rejection:
shielding and a differential signal path. The shield is good for
rejecting high frequency interference (i.e. RF), but is not as
effective at rejecting low frequencies (i.e. mains hum). The
differential signal path is good for rejecting low frequencies, but
not as effective at higher frequencies because the CMRR reduces with
increasing frequency. So, a good balanced connection relies on a
shield *and* a differential signal path to provide broadband
rejection of interference.
Shielding comes in three common forms: a braided shield, where the
wires are interleaved in a criss-cross pattern; a 'served' shield,
where the wires are spiral wound in the same direction; and a foil
shield where the inner conductors are wrapped in a plastic-backed
foil with a drain wire for soldering.
Braided shielding offers around 85% RF rejection (IIRC), but makes a
stiffer cable and is marginally harder to assemble (for what that's
worth). A served shield is more flexible and is easier to assemble,
and can match or exceed the rejection of a braided shield *if* in
good condition. A foil shield is very easy to assemble (due to the
drain wire) and can offer even higher rejection, but is best suited
to permanent installation work because it does not like being flexed.
As for the twisted pair inside the shield, it is important that the
physical geometry of the pair remains the same throughout the length
of the cable otherwise the rejection is compromised. For this reason,
cable manufacturers usually pack two strands of string or similar
into the cable, to fill in the space between the inner conductors and
the shield and thereby prevent the inner conductors from moving
outside of their required physical geometry. (IIRC, star quad designs
do not need this because the additional two inner conductors serve
the same purpose.)
As a further matter of interest, all cables have a specification
known as Minimum Bending Radius (MBR); this is the smallest radius
the cable can be bent or curved into before the internal geometry of
the twisted pair is compromised and/or significant gaps form in the
shielding. Exceeding the MBR will reduce the cable's rejection, and
usually this is permanent. It's like rolling or bending a piece of
paper - you can roll or bend it a certain amount and it returns to
normal, but if you roll or bend it tight enough to create a fold, it
will always be there. A permanent kink in the cable usually means the
MBR has been exceeded at that point, and things are no longer as good
as they could be - either the internal geometry has been disturbed,
the shield has been compromised, or both. Such kinks are often the
point where the cable will eventually fail. Many people roll up their
cables and then tie the end off in a nice tight knot, using the cable
itself! This is a disaster... it is much better to use velcro cable
straps or similar to tie off your cables.
As for connectors, gold-plated pins/sockets are worth having for any
kind of outdoor work because gold does not tarnish/oxidise and
therefore provides a good connection every time. Contrary to popular
belief, gold is not the best conductor, being equivalent to copper
IIRC. Silver is a superior conductor, but, like copper, tarnishes.
Finally, there is the choice of XLR shell. I have been privvy to
tests by a major microphone manufacturer whereby microphone cables
with different types of connector were placed in a heavy RF field and
the amount of rejection was measured. The best RF rejection came from
using the long-bodied Switchcraft connectors, with the shell
connected to the cable's shield.
Personally, I'm happy as a yak in a lush green pasture with Mogami
cable and Neutrik XLRs (black shells, gold pins).
- Greg Simmons
|