Preamplifiers and Mixers 739to achieve a DLP of 5° or less from 20 Hz to 20 kHz,
frequency response must extend from 0.87 Hz to
35 kHz, assuming 6 dB per octave (first order) filter
responses. If the low-pass filter is a second order Bes-
sel, the cutoff frequency can be as low as 25 kHz.^4
Notice that extreme high-frequency response is not
required, but extended low-frequency response is!
Phase distortion not only alters musical timbre, but it
has potentially serious system headroom implications as
well. Even though frequency response may be flat, peak
signal amplitudes can increase up to 15 dB after passing
through a network with high phase distortion. This can
be a serious problem in digital recording systems. Even
ultrasonic phase distortions caused by undamped reso-
nances can excite complex audible cross-modulation
products in subsequent nonlinear (any real world)
amplifier stages.^5
Low-frequency phase distortions are often described
as muddy bass and high-frequency phase distortions as
harshness or midrange smear. The complex cross-modu-
lation products are usually described as dirty sounding
and often are the cause of listener fatigue.
21.1.2.5 Common-Mode Rejection, Phantom Power,
and RF Immunity
Common-mode rejection, as discussed in Chapter 11,
Audio Transformers, is not just a function of the ampli-
fier input circuitry. It depends on the impedance balance
achieved by the combination of the microphone’s output
circuitry, cable, and the preamp’s input circuitry. Com-
mon-mode rejection ratio (CMRR) is seldom an issue
with dynamic microphones because, as shown in Fig.
21-1, the common-mode impedances are small parasitic
capacitances. However, when phantom power is
involved, very high CMRR can be difficult to achieve.
The circuitry in the microphone that extracts phantom
power from the two signal lines, as shown in the exam-
ples in Chapter 16, Microphones, can unbalance their
line impedances to ground. The resistors that supply
phantom power, shown in the preamplifier of Fig.
21-13, must also be tightly matched to achieve high
CMRR. For example, CMRR may be limited to 93 dB if
±0.1% resistors are used and may be limited to 73 dB if
±1% resistors are used. For comparison, the JT-16B
transformer used in Fig. 21-13 achieves a CMRR of
117 dB when phantom power resistors are absent.
Sometimes, as in Fig. 21-9, phantom power is supplied
through a center-tap on a microphone input trans-
former. This presents a transformer design problem that
can be even more difficult—simultaneously matching
both the number of turns and the dc winding resistance
on each side of the center tap.
RF interference, usually in the form of common-
mode voltage, is another potential problem for micro-
phone preamplifiers because it is likely to be demodu-
lated in amplifier circuitry. In transformer-less circuits,
suppression measures usually consist of capacitors from
each input to ground and sometimes series resistors,
chokes, or ferrite beads. Unless the capacitors are care-
fully matched, they will unbalance the common-mode
input impedances and degrade CMRR. Because they also
lower common-mode input impedances, they can make
the circuit more sensitive to normal impedance imbal-
ances in the microphone. These tradeoffs can be largely
avoided by using a Faraday-shielded input transformer
that has inherent RF suppression characteristics.
A good microphone preamplifier should also be free
of the so-called pin 1 problem. The microphone cable
should be free of shield-current-induced noise (SCIN),
which can be a serious problem with foil shield and
drain wire construction. Both of these problems are
discussed in Chapter 15.21.2 Real-World Preamp and Mixer Designs21.2.1 TransformersManufacturers of microphone preamplifiers have a nat-
ural desire to differentiate their product from all others.
One of the major divisive issues is the use of audio
transformers. According to the antitransformer camp,
all audio transformers have inherent limitations such as
limited bandwidth, high distortion, mediocre transient
response, and excessive phase distortion. Unfortu-
nately, many such transformers do exist and not all of
them are cheap. The makers of such transformers are
simply ignorant of sonic clarity issues, have a poor
understanding of the engineering tradeoffs involved, or
are willing to take manufacturing shortcuts that compro-
mise performance to meet a price.
As stated earlier, bandwidth and phase distortion are
intimately linked in any electronic device. A very high
level of performance can be reached with proper trans-
former design. Consider the Jensen JT-16B microphone
input transformer. Its frequency response is 3dB at
0.45 Hz and 220 kHz and 0.06 dB from 20 Hz–
20 kHz, with a second order Bessel high-frequency
roll-off characteristic. Low frequency roll-off is less
than 6 dB per octave owing to properties of the core
material, which further improves phase performance. Its
deviation from linear phase is under 2° from