Grounding and Interfacing 1197upper and lower arms have widely differing imped-
ances—e.g., when upper arms are very low and lower
arms are very high, or vice versa. Therefore, we can
minimize the sensitivity of a balanced system (bridge) to
impedance imbalances by making common-mode
impedances very low at one end of the line and very high
at the other. This condition is consistent with the require-
ments for voltage matching discussed in Section 32.3.2.
Most active line receivers, including the basic differ-
ential amplifier of Fig. 32-28, have common-mode
input impedances in the 5 k: to 50 k: range, which is
inadequate to maintain high CMRR with real-world
sources. With common-mode input impedances of
5k:, a source imbalance of only 1: , which could
arise from normal contact and wire resistance varia-
tions, can degrade CMRR by 50 dB. Under the same
conditions, the CMRR of a good input transformer
would be unaffected because of its 50 M:
common-mode input impedances. Fig. 32-29 shows
computed CMRR versus source imbalance for different
receiver common-mode input impedances. Thermal
noise and other limitations place a practical limit of
about 130 dB on most actual CMRR measurements.
How much imbalance is there in real-world signal
sources? Internal resistors and capacitors determine the
output impedance of a driver. In typical equipment, Zo/2
may range from 25 to 300:. Since the resistors are
commonly ±5% tolerance and the coupling capacitors
are ±20% at best, impedance imbalances up to about
20 : should be routinely expected. This defines a
real-world source. In a previous paper, this author has
examined balanced audio interfaces in some detail,
including performance comparisons of various receiver
types.^24 It was concluded that, regardless of their circuit
topology, popular active receivers can have very poor
CMRR when driven from such real-world sources. The
poor performance of these receivers is a direct result of
their low common-mode input impedances. If
common-mode input impedances are raised to about
50 M: , 94 dB of ground noise rejection is attained
from a completely unbalanced 1 k: source, which is
typical of consumer outputs. When common-mode
input impedances are sufficiently high, an input can be
considered truly universal, suitable for any source—
balanced or unbalanced. A receiver using either a good
input transformer or the InGenius® integrated circuit^25
will routinely achieve 90–100 dB of CMRR and remain
unaffected by typical real-world output imbalances.
The theory underlying balanced interfaces is widely
misunderstood by audio equipment designers. Pervasive
use of the simple differential amplifier as a balanced
line receiver is evidence of this. And, as if this weren’t
bad enough, some have attempted to improve it.
Measuring input X and Y input impedances of the
simple differential amplifier individually leads some
designers to alter its equal resistor values. However, as
shown in Fig. 32-30, if the impedances are properly
measured simultaneously, it becomes clear that nothing
is wrong. The fix grossly unbalances the common-mode
impedances, which destroys the interface CMRR for
any real-world source. This and other misguided
improvements completely ignore the importance of
common-mode input impedances.
The same misconceptions have also led to some
CMRR tests whose results give little or no indication of
how the tested device will actually behave in a
real-world system. Apparently, large numbers of
designers test the CMRR of receivers with the inputs
either shorted to each other or driven by a laboratory
precision signal source. The test result is both unreal-
istic and misleading. Inputs rated at 80 dB of CMRR
could easily deliver as little as 20 dB or 30 dB when
used in a real system. Regarding their previous test, theFigure 32-28. Basic differential amplifier.+ In- In
Differential amplifierOut+Figure 32-29. Noise rejection versus source imped-
ance/imbalance.Imbalance— 7CMRR—dB1m 10m 100m 1 10 100 1k0
10
20
30
40
50
60
70
80
90
100
110
120
130
140Zcm = 5 k7Zcm = 50 k7Zcm = 50 M7