Handbook for Sound Engineers

Preamplifiers and Mixers 737

To a greater or lesser degree, the frequency response
of any microphone will be affected by the load capaci-
tance of the connected cable and preamplifier as well as
the input impedance characteristics of the preamplifier.
Perhaps this is why the selection of microphones and
preamplifiers is such a subjective issue.

21.1.2 Some Considerations in Practical
Preamplifiers

Because many aspects of preamplifier circuit design and
the tradeoffs involved are discussed in Chapter 25, Sec-
tions 25.6 and 25.9, we will discuss only a few topics
here.

21.1.2.1 Gain and Headroom

Microphone preamplifiers commonly have maximum
voltage gains of about 60 dB to 80 dB and minimum
gains from 0 dB to 12 dB. A typical microphone, such
as the Shure SM57, will have an output of 1.9 mV or
–52 dBu with a 94 dB SPL acoustic input. For a very
high acoustic input of 134 dB SPL, its output would be
190 mV or –12 dBu. But a high-sensitivity microphone
such as the Sennheiser MKH-40 will have an output of
25 mV or –30 dBu at 9 dB SPL and 2.5 V or +10 dBu
at 134 dB SPL. Such high input levels can actually
require the preamplifier to have a loss (i.e., negative
gain) to produce usable line level output. Such high
input levels can also overload the preamp. Both prob-
lems are most commonly avoided with an input attenua-
tor or pad, typically of 20 dB. See Chapter 11 for a
discussion of the distortion and level handling charac-
teristics of audio transformers.

21.1.2.2 Input Impedance

As shown in Fig. 21-5, some input transformers have
input impedances that load the microphone and, as dis-
cussed in the preceding section, alter the response of the
system at frequency extremes. However, well-designed
transformers such as the Jensen JT-16B have substan-
tially flat input impedance as shown in Fig. 21-6.

21.1.2.3 Noise

The random motion of electrons in electrical conductors
creates a voltage variously called thermal noise, white
noise, or Johnson noise after its first observation by
J. B. Johnson of Bell Labs in 1927. Thermal noise volt-

age is proportional to both temperature and the resis-

tance of the conductor and is calculated as follows:^1

(21-1)

where,

Et is the thermal noise in rms volts,

k is Boltzmann’s constant or 1.38 × 10^23 Ws/°K,

T is the temperature of the conductor in degrees Kelvin,

R is the resistance of the conductor in ohms,

'f is the noise bandwidth in Hertz.

At a room temperature of 300°K (80°F or 27°C),

4 kT=1.66×10^20. For noise in the audio band of

20 Hz to 20 kHz, bandwidth is 19.98 kHz. It’s important

to note that noise bandwidth here refers to a rectangular

“brick wall” response, not the more conventional

measure at the 3 dB points. For a 150: resistance

under these conditions, noise is

For a 200: resistance under the same conditions, noise

is

Here we use the nominal impedance of an idealized

microphone simply to allow a simple but fair compari-

son of preamplifier noise performance.

Regardless of whether the conductor is copper wire,

silver wire, an expensive metal-film resistor, or a cheap

carbon resistor, the thermal noise is exactly the same!

Excess noise refers to additional noise generated when

dc flows in the resistor. Excess noise varies markedly

Figure 21-6. Input impedance of a Jensen JT-16B input

transformer.

3000

2750

2500

2250

2000

1750

1500

1250

1000

750

500

250

0

Jensen transformers JT-16-B Input Z (7  vs. frequency (Hz)

10 Hz 100 Hz 1 kHz 10 kHz 100 kHz

Frequency - Hz

Et= 4 kTR'f

223 nVrms –= 133.0 dBV

–= 130.8 dBu.

258 nVrms –= 131.8 dBV

–= 129.5 dBu.