Handbook for Sound Engineers

Grounding and Interfacing 1201

one point to keep it at ground potential and guard
against RF interference.26,27 Based on my own work,
there are two additional reasons that there should
always be a shield ground at the driver end of the cable,
whether the receiver end is grounded or not, see
Figs. 32-35 and 32-36. The first reason involves the
cable capacitances between each signal conductor and
shield, which are mismatched by 4% in typical cable. If
the shield is grounded at the receiver end, these capaci-
tances and driver common-mode output impedances,
often mismatched by 5% or more, form a pair of
low-pass filters for common-mode noise. The mismatch
in the filters converts a portion of common-mode noise
to differential signal. If the shield is connected only at
the driver, this mechanism does not exist. The second
reason involves the same capacitances working in
concert with signal asymmetry. If signals were perfectly
symmetrical and capacitances perfectly matched, the
capacitively coupled signal current in the shield would
be zero through cancellation. Imperfect symmetry
and/or capacitances will cause signal current in the
shield. This current should be returned directly to the
driver from which it came. If the shield is grounded at
the receiver, all or part of this current will return via an
undefined path that can induce crosstalk, distortion, or
oscillation.^28

With cables, too, there is a conflict between the star
and mesh grounding methods as discussed in Section
32.4.3. But this low-frequency versus high-frequency
conflict can be substantially resolved with a hybrid

approach involving grounding the receive end of cables

through an appropriate capacitance (shown in the third

device of Fig. 32-33).26,27 Capacitor values in the range

of 10 nF to 100 nF are most appropriate for the purpose.

Such capacitance has been integrated into the Neutrik

EMC series connectors. The merits of this scheme were

the subject of several years of debate in the Audio Engi-

neering Society Standards Committee working group

that developed AES48.

As discussed in Section 32.2.5, twisting essentially

places each conductor at the same average distance

from the source of a magnetic field and greatly reduces

differential pickup. Star quad cable reduces pickup even

further, typically by about 40 dB. But the downside is

that its capacitance is approximately double that of stan-

dard shielded twisted pair.

SCIN, or shield-current induced noise, may be one

consequence of connecting a shield at both ends. Think

of a shielded twisted pair as a transformer with the

shield acting as primary and each inner conductor acting

as a secondary winding, as shown in the cable model of

Fig. 32-37. Current flow in the shield produces a

magnetic field which then induces a voltage in each of

the inner conductors. If these voltages are identical, and

the interface is properly impedanc balanced, only a

common-mode voltage is produced that can be rejected

by the line receiver. However, subtle variations in phys-

ical construction of the cable can produce unequal

coupling in the two signal conductors. The difference

voltage, since it appears as signal to the receiver, results

in noise coupling. Test results on six commercial cable

types appear in reference.^29 In general, braided shields

perform better than foil shields and drain wires.

And, to make matters even worse, grounding the

shield of balanced interconnect cables at both ends also

excites the pin 1 problem if it exists. Although it might

appear that there’s little to recommend grounding at

both ends, it is a widely accepted practice. As you can

see, noise rejection in a real-world balanced interface

can be degraded by a number of subtle problems and

Figure 32-35. Shield grounded only at driver.

Figure 32-36. Shield grounded only at receiver.

Figure 32-37. Shield of a shielded twisted pair cable is

magnetically coupled to inner conductors.