1950, say the author of this story, "No longer are 'aerials' merely
required to transfer electromagnetic energy into space," in reference
to airborne platforms. Following great advancements in radio and
radar technology during World War II, great interest lied in
what would later become referred to as 'stealth' technology and
in secure communications. The transition of aircraft speeds into
the realm of supersonic also mandated that projections beyond the
main airframe outline be either eliminated or very much minimized.
The long cable aerials that stretched from the cockpit area to the
tip of the vertical fin, and the round direction finding antennas
hanging from below could not be accommodated at airspeeds above
about 300 knots. The aerodynamic drag would be excessive and
the forces would tear the antennas apart. Douglas Aircraft set up
one of the first antenna measurement laboratories specifically to
address those issues both for airborne and shipboard platforms.
Thanks to Terry W. for providing this
The Antenna Research Laboratory
By Joseph M. Boyer, Consulting Engineer
Douglas Aircraft's laboratory eliminates costly full-scale experiments
by using tiny replicas in solving complicated antenna design problems.
Fig. 1. Tiny crystal receiver shown in engineer's right
hand, is used to detect signals from model antenna. Plate
on side of model plane is removable to permit receivers
to be installed within the hollow fuselage of plane.
Never before in the history of radio has interest in the
antenna beam been at such a feverish pitch! No longer are "aerials"
merely required to transfer electromagnetic energy into space. Experts
today, working with surrealistic shapes of metal and plastic, are
molding radiated energy into the precisely shaped beams needed for
the varied classes of radar - for highly eavesdrop-proof communication
links, or even changing the beam's contour from instant-to-instant,
automatically following the boiling vagaries of the Heaviside layer.
The center of all such investigation is the antenna research
laboratory. Here antenna engineers work with worlds in miniature.
Out on the model antenna range of one such laboratory it is not
uncommon to see a complete scale replica of a television station:
the tiny buildings, the accurately-made antenna towers, even the
green rolling hills of the surrounding country. This Lilliputian
model slowly revolves on a turntable, a large horn-type radiator
some distance away "illuminating" it with microwave signals. The
miniature antennas of the model station detect such energy and feed
it back to high gain amplifiers in the laboratory. Thus, as the
model turns, automatic plotting instruments draw an accurate trace
of the radiation pattern of the station for later study. Such model
tests save costly cut-and-try procedures previously made on full
Even more important, in view of our
National Preparedness Program, is the investigation of aircraft
antennas. With aircraft now operating both near and beyond the speed
of sound, no object of any kind is permitted to project from the
sleek, polished metal skin to add parasitic drag. This requirement
is a death warrant for the numerous masts and wires which once were
draped lavishly over aircraft exteriors. In the high-pressure search
for distinctly new antenna types which may be faired flush into
the skin of a high speed airplane, several of the large airframe
manufacturers have aided the radio art immeasurably by taking the
lead in such research. In order to see, at first hand, the evolution
of a new antenna, a visit was made to the El Segundo, California
antenna laboratory of the Douglas Aircraft Company which pioneered
in this field. Here, work begins with the presentation of the Navy
specifications to the aircraft antenna designer.
call for a v.h.f. communications antenna. This unit is to be mounted
flush within the skin of a high speed carrier type fighter, yet
provide full 360-degree coverage about the horizon. When used for
transmitting, the antenna must produce most of its signal in a zone
approximately twenty degrees above and below the airplane. Efficiency
must be equal to the older type protruding antenna because airborne
power requirements are stringent. Finally, as if to complete the
designer's frustration, such an antenna must be capable of operating
from 300 to 5901 megacycles while remaining matched to
the coaxial transmission line feeding it. Specifically, it must
not exceed a voltage standing wave ratio of 2 to 1.
resourceful engineer begins a strenuous period of reading the available
technical literature, making rough preliminary calculations, and
weighing and discarding a number of configurations which come to
mind. In this process the crude pencil sketches which litter his
desk would be unrecognizable to prewar engineers. There is not a
sign of wires or porcelain insulators. One sketch may show a small
square portion of the metal skin isolated from the surrounding surface
and fed by a tapered funnel section of coaxial line. Or perhaps
a flat disc of polystyrene a foot or so in diameter is shown, excited
at its center by a sphere of silver designed to function as a wide-band
Fig. 2. Diagram shows position of model aircraft
and rotation axis for each of the three principal radiation pattern
made during pattern study.
Finally, the antenna
designer may feel he has what is needed. Before he makes a preliminary
shop drawing he must refine his design. This step involves extremely
complex calculations. For some such problems he must discard his
slide rule, set up the equations he wishes solved, and pass them
on to electronic or mechanical computing machines. Satisfied that
his "brain child" has a good chance of success, the engineer authorizes
the experimental shop to fabricate a full size antenna and pass
it on to the antenna laboratory for measurements.
Fig. 3. Typical aircraft antenna radiation pattern. The
pattern shown was photo. graphed on the screen of antenna
range cathode-ray "pattern painter." Magnetic deflection
coils move in synchronization with rotation of model under
study, tracing out an accurate polar diagram of antenna
signal variation around the plane model.
Fig. 4. Operating and recording position. Shown are the
v.h.f. and microwave transmitters, power supplies, and switching
panel. In front of operator is a pen recorder and the Douglas
cathode-ray "pattern painter." The "full moon" labeling
device is seen as the white window below the cathode-ray
Fig. 5. A coaxial slotted line in use. The slotted coaxial
line is used to measure the voltage standing wave ratio
of the prototype antenna. The radiator under test is mounted
on the outside metal surface of the wall, directly behind
the Hewlett Packard Voltage Standing Wave Ratio meter shown
The antenna laboratory
technician, highly-trained and experienced in this specialized field,
first may mount the prototype antenna upon a large ground plane.
This usually is a metal wall forming one side or the roof of the
laboratory building. In some cases the antenna may actually be mounted
into a full scale wire cloth mock-up of the aircraft itself. A precision
section of slotted coaxial transmission line (Fig. 5) is connected
in series with the antenna and a laboratory v.h.f. oscillator. Beginning
at one end of the frequency range to be covered by the antenna,
the technician makes measurements of the voltage standing wave ratio
in the transmission line. If the antenna is a perfect match there
will be no change in the measured voltage from one end of the transmission
line to the other. Such "flat" lines, however, are rarely encountered.
There usually is a small v.s.w.r, but it must be under the called-for
2:1 ratio. If the designer has done his job properly this condition
will be met over the entire frequency range desired. So far so good,
but more hurdles remain to be cleared.
Once more an order
goes to the experimental model shop: "Fabricate one 1/20th scale
model of the antenna for range pattern tests." The men who receive
this assignment are not ordinary machinists or metalsmiths. They
are, for the most part, former instrument makers used to working
with tiny precision parts under a powerful lens. They are fantastically
ingenious in devising ways of soldering and welding parts the size
of a pin head into place within complex assemblies, of bending and
twisting metal into shape while it is glowing in the flame of an
alcohol lamp. An idea of the difficulty of their job can be obtained
when it is realized that ordinary RG 8/U coaxial cable reduced to
1/20th scale is the size of store string. The inner conductor of
such cable is the diameter of a human hair, yet must be soldered
to the minute antenna without melting an extremely thin, easily-destroyed
polyethylene sheath which insulates the assembly. Upon completing
his exacting task, the model shop craftsman places the tiny antenna
into the metal skin of a previously prepared 1/20th scale model
of the aircraft in which it is intended to see service.
Radiation Pattern Measurements
the basic idea behind the use of the model antenna pattern range
is this: an aircraft operates far from the earth. The only environment
which affects the antenna on the airplane is the configuration of
the craft itself. Any attempts to measure radiation patterns on
a full size aircraft resting on the earth would be futile. Patterns
taken by means of flight tests are not only prohibitively expensive,
difficult to measure and interpret, but usually end in doubtful
results. However, by reducing the aircraft to 1/20th or 1/40th of
the full scale dimensions it is possible to mount it from 40 to
60 wavelengths from the ground. This can be done because the operating
frequencies must also be multiplied by 20 or 40 to keep in step
with the model dimension change. That such theory is correct, when
suitable precautions are taken, has been demonstrated conclusively.
The scaled-down model aircraft, complete with its test antenna,
is mounted upon a special dielectric tower, the base of which rests
on a motor-driven turntable. Within the hollow belly of the little
plane is a simple receiver usually consisting of an impedance matching
transformer and a silicon crystal detector or hot wire bolometer.
With the tower placed as many as 100 wavelengths from the
laboratory building, technicians energize a tunable Klystron transmitter
which excites a large horn type antenna projecting toward the model
through the wall of the laboratory. The transmitter's signal is
amplitude modulated by a square wave with a repetition rate of 1000
cycles. A square wave is needed to avoid frequency modulation of
the Klystron. Operating frequency is carefully adjusted to be 20
or 40 times the full scale point in the spectrum where the antenna
is intended to function.
Reference to Fig. 2 will make clear the patterns to be described.
The antenna specialist refers to such patterns as "cuts." The first
"cut" is made by slowly rotating the model so that every portion
of the plane's horizontal axis is exposed to the radio beam from
the laboratory transmitting horn antenna. The model on the tower
is then turned 90 degrees and again rotated by means of the turntable,
exposing its nose, belly, topside surface, and tail, to the beam.
Finally, a "cut" is made presenting the wingtips, belly, and topside
surface of the model. This triad of cuts - the horizontal, longitudinal
vertical, and transverse vertical, are fundamental in any pattern
investigation and quickly tell if the radiation pattern of a new
antenna is going to meet specifications. At least the three patterns
just described must be made at frequent intervals over the simulated
radio spectrum in which the antenna is going to operate. An antenna
may frequently have the desired radiation pattern at one end of
its frequency range and fail miserably at the opposite extreme.
Leaving the antenna designer for the moment with his problem
let us enter the laboratory building proper and investigate the
equipment used to study the radiation characteristics of antennas.
Several racks of audio amplifiers are the first instruments seen.
These are quite special items. There are preamplifiers capable of
boosting the few millivolts or so of signal received from the model
to about 10 or 20 volts. This piece of equipment is linear in response
and features a tuned feedback network which permits the amplifier
to operate with full gain only at 1000 cycles. All other signals
of random frequency and noise are sharply attenuated. The output
of the linear preamplifier drives a logarithmic amplifier which
is also sharply tuned to 1000 cycles. Logarithmic response is desired
so as to properly record variations in the model signal which may
extend over 50 decibels or more. To graphically present the radiation
pattern several different types of recorders are used.
The most common is a so-called polar recorder in which a pen is
driven by signal variations from the model through the use of a
servo-mechanism. In appearance this unit may resemble an automatic
phonograph record changer. A circular piece of polar graph paper
is placed upon its turntable and centered by means of a pinpoint
of light at the center. The paper edges are clamped down by means
of small Alnico magnets. Rotation of the recorder turntable is synchronized
by means of selsyns to turn in step with the model out on the pattern
range. When the model is rotated the servo-driven pen moves back
and forth on a radius, tracing out the pattern.
Fig. 6. General view of antenna model pattern range. A scale
model of the Douglas "Skyraider" is shown mounted on the
motor driven dielectric tower. The large electromagnetic
horn antenna to the right is being turned to change electric
polarization of signal to model. Smaller horn to the left
of the picture covers the three centimeter frequency range.
Fig. 7. Scale model aircraft and antenna shown in process
of construction. Craftsman in foreground solders a connection
in minute cavity type slot antenna. The 1/20th scale aircraft
model shown is of wood.
is a cathode-ray pattern painter illustrated in Fig. 4. This instrument
has several important advantages over the pen type recorders. One
of the most valuable is lack of mechanical inertia. There are occasions
when a radiation pattern being recorded varies from a deep null
to maximum signal intensity within a fraction of a degree of rotation.
Even for the slow speeds at which the model turns (3/4 to 1 r.p.m.)
this condition requires the pen to whip over the graph paper at
an exceptionally fast rate. The consequent lack of response and
"overshooting" of the pen distort such patterns.
is absent in the cathode-ray "pattern painter." Here the magnetic
deflection coils actually rotate about the neck of the cathode-ray
tube in synchronism with the model. Thus, as the signal intensity
changes the electron beam can follow the speediest variation with
no time lag, no error. When used for radiation pattern plotting
the screen (long persistence) of the tube is photographed on 35
mm. film for a permanent record (Fig. 3), Another fine feature of
the particular model developed at the Douglas laboratory is an edge-lighted
Lucite disc seen in the illustration mounted below the cathode-ray
tube. This disc is called the "full moon" because of its characteristic
of glowing with evenly distributed white light. All pertinent data
such as frequency, aircraft type, and description of the "cut,"
is typed on transparent gummed paper and this is then fixed over
the face of the "full moon." Easily photographed on the same film
as the pattern, such a screen label feature permits the laboratory
to obtain a very complete, foolproof record of work in progress.
must have transmitters available to cover enormous ranges of frequencies.
To see the reason why, let us assume that the full scale frequencies
of three antennas to be tested span the region 80 to 1600 megacycles.
Not only must oscillators be on hand for these exact frequencies
but, in addition, if the model range measurements are to be made
at 1/20th scale, r.f. generators are required for the simulated
range 1600 to 32,000 megacycles. Spanning such an expanse of radio
territory calls for an imposing collection of coaxial cavity, and
"butterfly" type oscillators, many, many Klystron tuners as well
as elaborate high-voltage regulated power supplies and frequency
measuring equipment of great accuracy. It is no wonder that antenna
engineers always ask for bar-gains in frequency coverage when shopping
for transmitters; otherwise such equipment would overflow the laboratory.
To cover the multitude of problems which trouble an antenna
specialist's slumber would be beyond the scope of this article.
Some of the especially serious ones, however, may be of interest.
The first and worst of these is spurious reflected signals. Exactly
the same problem is faced by television service technicians in the
form of "ghosts." The aircraft model itself is, of course, placed
carefully "in the clear." Any posts, buildings, fences, or personnel
in its vicinity would reflect signals into the model as if they
were secondary transmitters. Such reflections, depending upon their
instantaneous phase, either add or subtract in certain directions
from the true magnitude pattern of the model.
The real villain of this story, however, is the ground or platform
upon which the antenna laboratory rests. "Splash" from this source
is almost impossible to eliminate completely. Great care is exercised
in designing the large sectorial horn antennas which "illuminate"
the models so that just enough beam width with uniform phase front
is produced to cover the model with r.f. energy. Even though this
precaution lowers the magnitude of floor "splash" it does not completely
remove it. Sometimes low metal fences properly called detraction
edges are placed on the model range to deflect the "splash" signal
into a harmless area. Placing these fences for each frequency used
(and sometimes as many as 200 "cuts" are made on a single model)
is more of an art than a science.
Fig. 8. Close-up of 1 cm. transmitter and horn antenna.
A complete 30,000 mc. Klystron transmitter, cavity wavemeter,
and high gain horn radiator makes only a light handful of
Fig. 9. Slot antenna and cable. The size of a pair of 1/20th
scale slot antennas and miniature coaxial cable may be judged
by comparison with hand holding them.
Fig. 10. View of computer showing vacuum tube bays. Mathematician
inspects plug board which inserts problem into the 1285
vacuum tube electronic computer used to solve complex antenna
equations. Such machines are now routine tools in the search
for new antenna designs and antenna improvements.
Fig. 11. View of "feed" end of large horn type antenna.
Coaxial cable shown supplies v.h.f. energy to probe "feed"
for the large horn type "illuminating" antenna. Microwatts
are precious, and technician carefully adjusts the matching
stub for the maximum obtainable signal.
Another troublemaker is
the small coaxial cable which conveys the detected signal from the
model down the tower to the laboratory. This is, of course, a metallic
conductor of many wavelengths projecting from the model. Pattern
distortion will be introduced by this cable, and only highly experienced
personnel can minimize this difficulty by judicious placement of
the cable when setting the model up for a "cut." To overcome this
hazard some researchers have actually placed midget transmitters
inside the model aircraft. Battery power or an air-driven generator
energized by a high pressure hose are used, but the attendant cooling
problems and frequency drift due to lack of power supply regulation
makes this technique a last resort measure.
of distance in wavelengths at the operating frequency between the
model under test and the "illuminating" horn antenna poses, at times,
a nightmarish enigma for the antenna worker. In order that an accurate
radiation pattern be secured, the model aircraft must sometimes
be placed as many as 100 wavelengths from the laboratory antenna,
otherwise true "free space" conditions are not realized. Even at
the microwave frequencies 100 wavelengths may be a sizable distance
physically. Unfortunately, the power output of laboratory type Klystron
tubes is only about 200 milliwatts for the region up to about 8000
mc. and 40 to 50 mw. for frequencies above this. Sensitivity of
the simple receivers used in the models is quite low and, upon numerous
occasions, the model simply cannot be placed at the required distance
and still secure sufficient signal to record a pattern. The model
tower must then be moved into the so-called "near zone" region and
many hours spent in calculations and educated guesses in order to
replot the pattern to some degree of accuracy. The uninitiated invariably
suggest going to higher powered transmitters such as radar pulsed
units. When such suggestor, however, ponders over the problem of
building or purchasing the number of high power, room-size radars
needed to cover the frequencies called for, he soon realizes that
a fairly large warehouse would be needed to mount them for use.
Rejoining the antenna engineer it is found that his new
antenna has successfully completed its preliminary radiation pattern
tests. While mildly jubilant he must still subject his creation
to an investigation to determine its response to cross-polarized
signals. Also he must investigate what effect additional structures,
such as wing mounted rockets or bombs, have on its pattern. The
worried frown will remain on his brow for some time to come as he
follows the antenna through the intricate maze of production decisions,
cost analysis, and lastly the flight test which places the final
stamp of approval on his work.
While emphasis has been placed
on the aircraft antenna because of its present importance, it should
be made clear that the laboratories of such institutions as Ohio
State University are carrying on programs of investigation into
many other aspects of the antenna problem. For example, in the study
of land-based radio stations scientists must content themselves
not only with dimensional perfection with regard to towers and buildings,
but must also actually design the soil of these scale models to
have proper conductivity at the higher model frequencies in order
that it simulate the soil found in the region under study. The "guess
and by-gosh" methods of the past in making costly antenna installations
are slowly giving way to exact knowledge.
Last but by no
means least, Naval research centers are engaged in measuring the
radiation patterns of antennas mounted within the complex maze produced
by a ship's masts, cables, and other marine structure. As might
be expected, Naval antenna designers must take the sea into account
when making their scaled-down ships for range tests. To an electromagnetic
wave a ship resting upon the sea appears to have an exact mirror
image directly beneath it. This can be duplicated on "the Naval
antenna model range by cutting a ship model off at its water line
and resting it upon a large sheet of metal. In lieu of this, two
ship models are constructed, sawed off at the water line, and one
fastened upside down to the waterline of the other. The technique
of making the actual radiation pattern measurements is identical
to that described for aircraft.
1 Military frequencies are classified. Those
given are only representative.