Handbook for Sound Engineers

1078 Chapter 28

28.4.3 Bias and Erase Circuits

The high-frequency signals required for biasing and
erasing all tracks of the tape are derived from a single
master oscillator so that no interference or beating of
multiple oscillators will occur. Older designs generally
employ a tuned push-pull multivibrator oscillator;
newer designs favor crystal-stabilized oscillators
utilizing digital circuitry. Several designs have used
separate bias and erase frequencies, with the erase
circuit running at one-third the bias frequency to mini-
mize the power dissipation on the erase head.

In all cases, the primary consideration is purity of the
bias and erase current waveforms. Any even-order
harmonics, including dc, second harmonics, fourth
harmonics, and so on, will create a detrimental rise in
the background tape noise, reducing the available SNR
for the recorder. Older designs, such as Fig. 28-44A,
relied heavily on push-pull circuits with balancing
transformers to minimize these even-order components.
Newer designs, such as Fig. 28-44B, favor filtering and
feedback control to reduce unwanted components. The
divide-by-two flip-flop eliminates any even-order
distortion in the oscillator waveform.

The erase head is typically coupled to the erase
source with an adjustable series resonating capacitor to
minimize the voltage required from the driver and to
filter out even-order components. A current sampling
resistor is frequently provided in the ground leg of the
erase head circuit so that the amplitude of the erase
current can be conveniently monitored.

28.4.4 Noise Reduction Systems

The SNR of an analog audio recorder is usually taken as

the difference between the residual biased tape noise

level and the level which produces 3% third harmonic

distortion at 1 kHz. In the ideal case this ratio is limited

by the tape speed, track width, and tape type. Once these

parameters are set, the maximum SNR is determined.

Direct analog noise reduction systems rely on the

masking effect of human hearing. If both a background

noise and a louder desired signal exist within the same

frequency band, the noise will be masked by the desired

signal. If, on the other hand, the noise and signals are in

different parts of the audio spectrum, such as a bass

guitar and high-frequency tape hiss, the noise will not

be masked. The perceived noise can be reduced if the

SNR is compromised during masking situations so that

unmasked noise can be reduced. This requires dynamic

change of the gain or transfer function of the system

depending on the program content.

Dolby™ and dbx™ noise reduction systems are

examples of amplitude-only encode/decode systems.

Both systems modify the amplitude of the signal to

squeeze the dynamic range of the input signal into a

smaller dynamic range that will avoid the noise and

distortion limitations of the recording tape. The fidelity

of these compander (compression/expander) systems is

limited not only by the tracking of the encode and

decode circuits, but also by the nature of the errors that

are generated by noise, nonlinearities, and frequency

response anomalies introduced by the record/playback

cycle of the tape recorder. These parasitic errors can

cause dynamic mistracking that will create distortions

of dynamic signals that may not be evident during

sine-wave testing.

The Dolby systems process the low-level signals by

boosting them during recording and then attenuating

them on playback. The original professional Dolby

system (Dolby A) subdivides the audio spectrum into

bands that are processed individually to optimize the

masking effect.

A later development, the Dolby SR system also adds

adaptive filters that change their cutoff frequencies as

the signal content varies. When the program material

includes information at high frequencies, the filter

opens up to full bandwidth. If no high-frequency

content is present, the cutoff frequency of the filter

slides down to match the program material. The filter

must be “intelligent” enough to distinguish between

desired audio signals and unwanted noise.

The Dolby SR system was quickly embraced by the

music and film markets as a method of raising analog

Figure 28-44. Typical bias and erase sources.

From

push-pull

oscillator

Vcc To record

head

To erase

head

A. Push-pull circuit with balancing transformers.

Crystal

reference

B. Filtering and feedback control.

To buffer

amp