December 1947 Radio-Craft
Wax nostalgic about and learn from the history of early electronics.
See articles from Radio-Craft,
published 1929 - 1953. All copyrights are hereby acknowledged.
In most instances the method and materials have changed over the
years, but fundamental principles of writing and reading data to
and from magnetic media are the same today as when this article
was written in 1947. If you find that your lexicon of technical
jargon lacks terms such as
remanence, then you might want to invest a few moments reading
this short article that appeared in Radio-Craft. I realize most
regular RF Cafe visitors won't be interested, but hopefully someone,
somewhere, searching for this information will now be able to find
it. Thanks for your indulgence.
Magnetic Recording - Recorder Design
Part III of a series, the first two parts
of which were headed Magnetism. This part deals with recorder design.
Next will be construction of a practical tape recorder.
By A.C. Shaney*
*Chief Engineer, Amplifier Corp. of America.
Fig. 1 - The Mail-A-Voice uses magnetic discs. Courtesy
Brush Development Corp.
Fig. 2 - This machine records 12,500 cycles. Courtesy
Amplifier Corp. of America
One of the most important elements in any magnetic recording
process is the magnetic carrier (the material upon which the magnetic
modulations are impressed or recorded). Magnetic carriers which
have been used successfully to date include metallic ribbon and
wire, nonmagnetic ribbon and wire plated with special magnetic coatings,
and magnetically coated and impregnated paper and plastic tapes.
Early experimenters successfully applied magnetic modulation to
a magnetically coated cylinder. A magnetic recorder and play-back
device which is now available (Fig. 1) utilizes magnetically coated
thin paper discs. This list is merely suggestive and by no means
exhausts all the possibilities. For example, it should be both economical
and practical to coat or impregnate magnetically plastic or cotton
fiber thread or other artificial fiber.
The actual magnetic properties of the magnetic carrier determine,
to a considerable degree, the overall performance characteristics
of the recording and playback system. They also play an important
determinative factor in the design of all other essential elements
and auxiliary components used in both the magnetic recording and
Many early and highly qualified technicians experimenting with
magnetic recording failed to evaluate properly all effects which
were normal functions of carrier characteristics, and as a result
came to erroneous conclusions. For example" it was believed (during
1932, that it was necessary to run a magnetic carrier at a speed
of 393.7 inches per second to obtain a frequency response up to
5,000 cycles. As a result, it was fallaciously concluded that the
necessity for the high speed "restricts the application of this
recording method to reproduction in the speech range." Today in
contrast, a commercially available unit (Fig. 2) is capable of recording
and reproducing up to 5,000 cycles with a carrier speed of 4 inches
per second - a reduction in carrier speed of nearly 100 times without
affecting frequency response, attaining, at the same time, many
other desirable improvements in noise reduction, increased dynamic
range, and lower distortion! Nine thousand cycles has been recorded
and played back at a carrier speed of 7 1/2 inches per second, and
12,500 cycles has been attained at a tape speed of 15 inches per
second. (Magnetically coated paper tape was used to attain the indicated
The rate of progress in the art of magnetic recording can be
measured by the carrier speed of the magnetic medium employed. Slow
speeds (with desired frequency response) indicate greater economy
of the magnetic medium and longer playing time for a given length
and cost of material
Progress in this respect has been little short of phenomenal.
For example, in 1932 118,000 feet of wire was required for a 1-hour
program reproducing up to 5,000 cycles. In 1943 the same program
quality and duration could be maintained with 11,000 feet of wire.
By 1946, it took only 1,250 feet of coated tape to duplicate the
All students of radio have been correctly impressed with the
idea that the rate of radio wave propagation (radio carrier speed)
is constant (300,000,000 meters per second). As a result, it is
simple to calculate wave length - the distance through which current
will travel within one cycle - from the following well-known formula.
λ = k/f
when λ = wavelength, f = frequency, and k = carrier speed
(186,000 miles per second or 300,000,000 meters per second).
In magnetic recording, the carrier speed is not a fixed and unvarying
constant. In fact, as previously explained, there is a strong tendency
to continually decrease its speed without affecting over-all response.
As a result, the magnetically recorded wavelength of a given frequency
will be a function of the magnetic carrier speed and can be found
from this simple expression:
λ = s/f
when s = magnetic carrier lineal speed.
Thus a 5,000-cycle signal when magnetically recorded on a tape
running 7 1/2 inches per second will have a wave length of 1.5 thousandths
of an inch. Similarly, a 5,000-cycle signal recorded at a carrier
speed of 4 inches will have a wave length of 8 ten-thousandths of
When it is realized that each complete wave length has 4 phases
(see Fig. 3), evidently each phase of 8 ten-thousandths of an inch
wave length will be only 2 ten-thousandths of an inch long (approximately
5 microns). We begin to get an idea of the minute dimensions involved
in attempting high-frequency magnetic recording on slow-moving media.
Carrier speed stability
Fig. 3-a - Four phases of sine wave. 3-b - Enlarged cross
section of paper tape to show internal and external magnetic
A casual examination of the formula (2) relating wave length
to speed and frequency indicates that if the magnetic wave length
is constant (and it is, if a fixed frequency has been recorded at
a constant speed) the reproduced frequency will be directly proportional
to lineal speed. If the lineal speed of the magnetic medium should
vary for any reason whatsoever, the reproduced frequency will similarly
vary. This produces a noticeable variation of frequency when sustained
tones are reproduced and resembles the turntable "wow" common to
disc recording and reproducing systems which do not employ absolutely
constant-speed turntables. To minimize any instantaneous speed variations,
it is important to avoid eccentricities in any of the driving members
involved in pulling the magnetic medium past the playback head.
A correctly designed flywheel should be used to smooth out cyclic
pulls common to "constant speed" synchronous "motors. A correctly
designed capstan drive and flywheel is illustrated in Fig. 4.
The load applied to the driving motor should be constant. Bent
reels or spools which scrape either the recording medium or adjacent
surfaces are the most common cause of frequency variations in a
properly designed mechanism. An idea of the desired constancy of
linear speed may be gained from the already established data which
indicates that an average listener can detect frequency deviations
in the order of 3/10 of 1% (3 parts in 1,000) within the frequency
range of 400 to 5,000 cycles. Therefore, instantaneous lineal speed
variations should be no greater than ± 0.1% (which allows a variation
of 2 parts in 1,000).
Magnetic carrier compliance
Early experimenters who attempted to increase the high-frequency
response of magnetic recording systems by increasing wire speeds
were bothered by pronounced variations in signal level mainly caused
by wire "flutter."
A stiff wire travelling at high speed will tend to assume some
natural period of vibration dependent upon its thickness, tautness,
and the distance between its supports. Flutter (a transverse vibration)
produces a minute frequency variation "wow" (because a slight change
in lineal speed takes place), but more important, it introduces
a varying pressure against the recording head which in turn changes
the air gaps between the magnetic medium and its pickup head. These
minute variations produce appreciable high-frequency level fluctuations
characteristic of flutter. Steel tapes travelling at high speeds
are similarly afflicted. Paper and plastic tapes, - because of their
increased compliance and slow speeds - are more easily passed by
the pickup head with negligible flutter effects. On the other hand,
their increased compliance makes it necessary to use pressure fingers
to keep the tape pressed, at a relatively fixed pressure, against
the recording and playback heads.
When discussing lineal speed stability and its relation to "wow,"
it was assumed that a magnetically recorded fixed-frequency signal
would maintain a fixed wave length. This is true as long as the
wire or tape doesn't stretch. If stretching does take place, because
of temperature and humidity changes between the recording and playback
process (and it may, because of the tensions or pull applied to
the medium during recording, rewinding, and playback), we then have
a new variable to consider.
A simple transposition of formula (2) produces
s = λf (3)
which indicates that if the speed remains constant, the reproduced
frequency is inversely proportional to wave length. In other words,
if the magnetic medium stretches the wave length increases and reproduced
frequency decreases! This effect will not be noticeable as long
as the "stretch" or dimensional stability is within ± 0.1%. Plastic
tapes and thin wires naturally will have some tendency to stretch.
The paper base used for coated tape is made of carefully selected
material and treated by prestretching for improved dimensional stability.
Fig. 4 - A correctly designed "tape puller" for magnetic
recording and playback devices.
As expected, one of the most important elements in magnetic recording
and playback is the actual magnetic and physical properties of the
magnetic carrier. In a magnetic coated medium, the coating itself,
and not the base, plays the most important role in the process.
Some of the more critical factors which determine the efficiency,
noise, response, constancy of output, overload characteristics,
and velocity of the carrier include it coercivity, remanence, particle
size, binder, dispersion, chemical composition, surface smoothness,
and coating thickness.
Each of these characteristics has a profound influence upon the
over-all characteristics and performance of the recording and playback
A brief. discussion of these factors together with the design
details of a suitable magnetic recording amplifier will be covered
in the next issue.
Two transpositional errors occurred in the October installment
of this series. In comparing the conductivity of electric and magnetic
circuits, the formulae should have been:
G = I/E
μ = B/H
G = conductance in mhos
E = e.m.f. (volts)
I = current (amps)
μ = permeability
H = magnetomotive force (oersteds)
B = magnetic flux (gausses)
Posted September 9, 2014